Download Extremely Low Frequency (ELF) Radio Wave Propagation: A review

Document related concepts

Aharonov–Bohm effect wikipedia , lookup

Time in physics wikipedia , lookup

Transcript
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Extremely Low Frequency (ELF) Radio Wave Propagation: A review
A.P. Nikolayenko1, A. Shvets1, and M. Hayakawa2
1
Usikov Institute for Radio-Physics and Electronics, National Academy of Sciences of the Ukraine, Kharkov,
Proskura st. 12, 61085, Ukraine
2) The University of Electro-Communications, Advanced Wireless & Communications research Center, 1-5-1
Chofugaoka, Chofu Tokyo 182-8585, Japan
Abstract.
The Extremely Low Frequencies (ELF) extend from 3 Hz to 3 kHz. This formal limit
corresponds to real physical phenomena in subionospheric radio propagation: the Schumann
Resonance (SR) observed in the band between 4 Hz and 40 Hz and the transverse resonance
having the basic frequency about 1.7 kHz. When speaking about ELF radio waves, we usually
have in mind the subionospheric propagation when electromagnetic wave travels in the
spherical dielectric stratum formed by the Earth‟s surface and the lower edge of the ionospheric
plasma. The dielectric shell has a small thickness of about 60–100 km in comparison with the
Earth‟s radius of ~6366 km. This non-conducting layer forms a spherical cavity resonator, with
two sets of eigen-values. One of them corresponds to the global electromagnetic resonance
predicted by Schumann [1952], which is often regarded as SR. Its peak frequencies
(frequencies of maxima in the power spectra of natural ELF radio noise) are approximately 8,
14, 20, Hz, etc. At such low frequencies, only transverse electromagnetic (TEM) wave
propagates in the cavity having the non-zero vertical electric Er and horizontal magnetic H
field components. If we account for the finite ground conductivity, a small longitudinal electric
field E appears. It is clear that we use the spherical polar coordinate system (r, , ) with the
origin at the Earth‟s center and the polar  = 0 axis directed to the vertical electric dipole
source.
Owing to small losses and a great wavelength, the resonance signals can multiply & circle
the globe, so that direct and antipodal waves interfere forming a set of maxima in the power
spectrum of vertical electric or horizontal magnetic fields positioned at the above-mentioned
frequencies. Lightning strokes of the global thunderstorms occurring at the rate of 50–100
discharges per second serve as the source of natural ELF radio noise. SR might be regarded as
“longitudinal” one: it occurs in the waves propagating along the Earth‟s surface.
The second type of resonance corresponds to vertically propagating radio waves, therefore, it
is called the transverse resonance. Relevant waves are captured between the ground and
ionosphere, so that their basic frequency depends on the ionosphere height and is equal to 1.7
Copy right @ESES
IJEAR
1
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
kHz when the thickness of the ground–ionosphere gap is 88 km. Transverse resonance is a local
phenomenon, because it is not linked with the round-the-world propagation. In radio
engineering, the eigen-values of transverse resonance are regarded as cutoff frequencies of the
modes propagating in the Earth–ionosphere duct. The basic frequency is the cut-off frequency
of the first-order mode, the second-order mode has the cutoff equal to the doubled basic
frequency of transverse resonance, etc. Obviously, only the basic transverse resonance
frequency definitely belongs to the ELF band, while the higher-order modes are found within
the very low frequency (VLF) band.
Key words: 1 ; ELF(extremely low frequency),2; ELF propagation theory, 3; atmospheric
conductivity, 4;VLF(very low frequency), 5;Schumann resonance (SR), 6;ELF transients, 7; Qbursts, 8;global thunderstorm activity, 9;global warming, 10; terminator effect, 11;space weather,
12; inverse problem, 13;transverse resonance,14; tweeks.
1.
Introduction
We describe the Extremely Low Frequency (ELF) radio propagation in the Earth-ionosphere cavity.
The formal description of electromagnetic field is presented in the frequency and in the time
domains. The expected spectra are demonstrated of the global electromagnetic (Schumann)
resonance. The waveforms are given of the ELF transients or Q-bursts. Models of ELF propagation
constant are listed together with the efficient approximate models of the lower ionospheric
conductivity. The vertical profiles of conductivity are given that provide realistic Schumann
Resonance (SR) data in the direct field computations, such as FDTD (Finite Difference Time
Domain) method. Observational aspects are addressed including a brief description of the field
sources, instrumentation, and typical signal processing with illustrations of actual data. Information
is presented deduced from the ELF observations, e.g., diurnal/seasonal variations of global
thunderstorm activity. Impact of ionospheric non-uniformities is mentioned in the context of its
separation from alterations in the global thunderstorm activity. A comparison is made of the model
time domain pulse with the unique observations of distant Q-bursts. Transverse resonance in the
Earth–Ionosphere cavity is introduced as observed as tweek atmospherics. A solution of relevant
inverse problem is briefly described allowing simultaneous inference of the source distance and
parameters of the lower ionosphere (“Kharkov technique”). The other topics are listed that could
not be included in the article of limited size.
Copy right @ESES
IJEAR
2
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Figure 1: Two types of electromagnetic resonance (longitudinal (Schumann) and transverse) in the Earth –
ionosphere cavity
The ELF extends from 3 Hz to 3 kHz. This formal limit corresponds to real physical phenomena
in subionospheric radio propagation: the SR observed in the band between 4 and 40 Hz and the
transverse resonance with the basic frequency about 1.7 kHz (see Figure. 1). When speaking about
ELF radio waves, we usually have in mind the subionospheric propagation when electromagnetic
wave travels in the spherical dielectric strata formed by the Earth‟s surface and the lower edge of
the ionospheric plasma. The dielectric shell has a small thickness of about 60–100 km in
comparison with the Earth‟s radius of ~6366 km. This non-conducting layer forms a spherical
cavity resonator, having two sets of eigen-values. One of them is the global electromagnetic
resonance predicted by Schumann [1952] which is often regarded as SR. Its peak frequencies are
approximately 8, 14, 20, Hz, etc. At such low frequencies, only transverse electromagnetic (TEM)
wave can propagate in the cavity having the non-zero vertical electric Er and horizontal magnetic
H field components. If we account for the finite ground conductivity, there appears a small
longitudinal electric field E . It is clear that we use the spherical polar coordinate system (r, , )
with the origin at the Earth‟s center and the polar  = 0 axis directed to the vertical electric dipole
source.
Copy right @ESES
IJEAR
3
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Because of small losses and great wavelength, resonance signals can multiply & circle the globe,
so that direct and antipodal waves interfere forming a set of maxima in the power spectrum of
vertical electric or horizontal magnetic fields positioned at the above-mentioned frequencies.
Lightning strokes of the global thunderstorm activity occurring at the rate of 50–100 discharges in a
second serve as the source of natural ELF radio noise. SR might be regarded as “longitudinal”: it
occurs due to waves propagating along the ground.
The second type of resonance corresponds to radio waves propagating vertically, so it is called the
transverse resonance. Relevant waves propagating vertically are trapped between the ground and
ionosphere, so that their basic frequency depends on the ionospheric height and is equal to 1.7 kHz
when the thickness of the ground – ionosphere gap is 88 km. Transverse resonance is a local
phenomenon, because it is not linked with the round-the-world propagation. On the other hand,
frequencies of transverse resonance are regarded as the cut-off frequencies of modes propagating in
the Earth–ionosphere duct. The basic frequency is the cutoff frequency of the first mode, and the
second mode has the cutoff equal to the doubled basic frequency of transverse resonance, etc.
Obviously, only the basic transverse resonance definitely belongs to the ELF band and higher-order
modes are found within the very low frequency (VLF) band.
2.
Radio Propagation at ELF
2.1. Radio propagation theory
In the simplest approach one considers the Earth–Ionosphere cavity formed by two perfectly
conducting concentric spheres with the cavity height h much smaller than the Earth radius a. The
Schumann‟s formula is derived for the eigen-frequencies f n 
c
nn  1
2a
where n = 1, 2, … is the
SR mode number [Schumann, 1952]. After substituting the light velocity c and the Earth‟s
circumference 2a, one obtains the equation fn = 7.5[n(n+1)]½ and the following succession of the
eigen-values: 10.6, 18.4, 26.0, 33.5 Hz, etc. Peak frequencies fPEAK of SR are observed in the power
spectra of natural ELF radio noise being about 8, 14, 20, 26 Hz, etc. These also vary with the mode
number as [n(n+1)]½, and only the factor 7.5 Hz should be replaced with 5.7 Hz (see Figure 2).
Copy right @ESES
IJEAR
4
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Figure 2: Eigen-values of perfect and observed peak frequencies of the Earth–ionosphere cavity
Deviations of observed peak frequencies from the [n(n+1)]½ dependence do not exceed 3%, and
this proves that the Earth is a sphere. Plots in Figure. 2 show the congruent curves of the perfect
and real cavities, and the both deviate from the linear asymptotic dependence relevant to Cartesian
geometry (see Chapter 2 in [Nickolaenko and Hayakawa 2002] for details).
Formulas for the electric and magnetic fields of the cavity are obtained from the solution of the
boundary problem, and the detailed description might be found in [Nickolaenko and Hayakawa 2002;
Wait, 1970; Galejs, 1972; Bliokh, 1980 ; Nickolaenko and Hayakawa 2014]. A single zero-order
mode propagates in the cavity at frequencies below 1.6–1.7 kHz, and it is often regarded as TEM
mode. Only two field components might be detected on the perfectly conducting ground:
Er (  ) 
M C (  ) i   1 P cos    

sin
4ha 2
H    
M C   P1cos    
4ha
sin
(1)
(2)
Here Er and H are the vertical electric and horizontal magnetic fields in V/m and A/m
correspondingly,  is the angular frequency, MC is the current moment of the source (Am), h is the
effective height of the cavity in m, a is the Earth‟s radius in m,  is the permittivity of free space in
F/m,  is the dimensionless propagation constant,  is the source–observer angular distance in
Copy right @ESES
IJEAR
5
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
radians, and P(x) and P1(x) are the Legendre and associated Legendre functions of order  and
argument x. The vertical electric dipole source of the current moment MC is positioned at the pole  =
0 of the spherical polar coordinate system (r, , ) with the origin at the Earth‟s center.
The physical meaning of the parameter () = 1()i2() becomes clear at high frequencies
where |  |ka>>1 and the asymptotic expansions are valid for the Legendre functions [Gradstein and
Ryzhik ,1963]:
P cos     
2
cos  1 2      4
   1 2sin
2   1 2
sin  1 2      4
sin
P 1cos     
(3)
(4)
Here  should not be equal to 0 or . At high frequencies the imaginary part 2() responsible for the
wave attenuation is large, so that the trigonometric functions are replaced by exponential functions:
P cos    
2
i
exp i1  1 2   4  2 
sin
   1 2sin
(5)
P 1cos    
2  1 2
i
exp i1  1 2   4  2 
sin
 sin
(6)
The above formulas suggest that the time dependence is of the form exp(+it), and so the field
components in Eqs. (5) and (6) are proportional to exp [(it1)]. Hence Eqs. (5) and (6) contain
the direct wave traveling from the source to the observer along the great circle path, and () is the
propagation constant when the path is measured in radians. Formulas (3) and (4) include both the
direct and antipodal waves. The antipodal wave travels along the same great circle, but reaches the
observer via the source antipode and has smaller amplitude. At SR frequencies we have |  |  1, so
we cannot use the asymptotic expansions (3–6) corresponding to |  | >> 1. The zonal harmonic
expansion is used here for the Legendre functions.
SR occurs physically when the wavelength is comparable with the Earth‟s circumference, direct
and antipodal waves arrive at the observer having almost the same amplitudes, and the observed
field strongly depends on the wave interference. The amplitude increases at the frequencies where
the waves meet in phase, and we speak about the SR maxima. At frequencies below 100 Hz, the
zonal harmonic series representation (ZHSR) is used [Jones, 1969; Jones, 1970]:
Er (  )  
Copy right @ESES
M C (  ) i   1
4ha
2


( 2n  1 )P ( cos )
 nn  1 n   1
(7)
n 0
IJEAR
6
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
 2n  1
M C   
P1  cos  

4 ha n1 n n  1     1 n
H   
(8)
We conclude that the following parameters are necessary in the computations of expected ELF
fields: the source current moment (MC), the effective height of the Earth–Ionosphere waveguide
(h), the source–observer distance (SOD) (θ), and the propagation constant (). The source current
moment MC(f) is supposed to be a constant equal to 108 A m Hz½, which corresponds to a realistic
lightning stroke with the channel length of 4 km and the current spectral density of 25 kA Hz½.
Such a current moment corresponds to a delta-function in the time domain, and the relevant charge
moment is a step function. The effective height of the cavity and the propagation constant will be
discussed below.
Each term in the ZHSR (7) and (8) is called the mode, and it contains the resonance denominator
that might be written in the form: n(n+1)  (+1) = (n  )(n++1). The ELF propagation constant
(f) is a complex function of frequency, and the following condition is satisfied at SR: Re [(fPEAK)] 
n, so that the resonance denominator becomes small and the n-th term in Eqs. (7) and (8) is large.
Thus the resonance peaks appear in ELF spectra. Each resonance term of ZHSR is multiplied by the
relevant Legendre or associated Legendre polynomials, and the products sum provides the unique
outline of SR amplitude spectra for a given SOD  as seen in Figure 3.
The series in Eqs. (7) and (8) converge poorly, because this is conditioned by the field singularity
at the source point. The method of acceleration convergence of the series is adopted, and we show the
final result (see [Nickolaenko and Hayakawa 2002; Nickolaenko and Hayakawa 2014] for details and
references).
 au
P  x 
 
 R   b Pn x 
sin
n 0 a
where R  2 R1  R2  2   1  1R3  33   1  2R4 , R1  ln



(9)
1 x  2
, R2  1  1  x R1  21  x  ,
1 x

R4   R1 5x  1x  12 12  15x 2  27 x  14 36  5 1  x 2 21  x  12  11 21  x 3 18 ,


R3  1  3R1 1  x 3x  1 4   x  21  x 3 3 4 , a u  2n  2   12  22   1  12  17 2   12  74   1  24 ,


a b  n  n   1n  1n  2n  3n  4 .
Similar expressions are valid for accelerated computations of associated Legendre function:
Copy right @ESES
IJEAR
7
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
 bu

P1  x  2   1
 

2 R3  9 R4   b Pn x 


sin
n 0 b
1  x2 

(10)
where bb     1  n  1n   1  n  1n  2n  1n  2n  3n  4 and



bu  4n3   1  12 2n2 19   1  78 n 2 2   12  19   1  66  17 2   12  34   1  24 .
SR over the frequency – distance plane
2.11
Field expansions with accelerated convergence allow us to compute the SR spectra for any SOD.
We plot the 2D (2 dimensional) amplitude distributions in Figure 3 of electric and magnetic fields
over the frequency–distance plane. Frequency is shown along the abscissa on logarithmic scale, and
SOD is shown on the ordinate in mega meters (1 Mm= 1,000 km). Field amplitude is shown with
color. The left graph corresponds to the vertical electric field and the right one depicts the amplitude
of the horizontal magnetic field component.
We had used the heuristic linear dependence of propagation constant (see below). The source
current moment was equal to 108 A m s, and the fields were measured in mV m Hz½ and A m Hz½
correspondingly. The maps illustrate the interaction of direct and antipodal waves that forms the
resonance patterns in the space and frequency. A broad spectral maximum occupies the frequencies
of some hundred Hertz when the SOD is small. This is the area where the slow-tail ELF
atmospherics might be observed. With an increase in SOD, the wave attenuation also increases and
the spectral peaks rapidly decrease [Nickolaenko and Hayakawa 2002; Nickolaenko and Hayakawa
2014].
Figure 3: Amplitudes of E and H fields over the frequency–distance plane, where 1 Mm=1,000km
Copy right @ESES
IJEAR
8
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Figure 3 includes a system of summits (shown by red color) corresponding to individual SR
modes. Amplitude of the vertical electric field grows toward the source and to its antipode where the
well-resolved SR peaks are visible. These two „highlands‟ are divided by a wide „valley‟ centered
around the distance of 10 Mm. One observes a series of peaks extended vertically in the maps, i.e.,
along the SR modal frequencies. A few summits and saddles are positioned here corresponding to
maxima and to the nodes of a given resonant mode, which are described by the corresponding
Legendre polynomial. The central valley gradually descends toward the higher frequencies owing to
an increase in the wave attenuation.
The side canyons are also present in Figure 3 extended from lower left to upper right. These
appear due to the interaction of direct and antipodal waves and tend to connect the nodes of
adjacent modes. These areas correspond to the condition: (f) ()  const describing the families
of hyperbolic curves. There exists a secondary hardly visible symmetric system of canyons
approaching the source.
The right map in Figure 3 refers to the horizontal magnetic field. Interaction of direct and
antipodal waves is also present here. Both the fields contain a broad slow-tail peak when the SOD is
small. A succession of equidistant saddles and peaks is seen at a fixed distance in both maps. The
„antipodal effect‟ becomes visible at high frequencies: a series of depressions here tend to become
extended horizontally.
There are distinctions between the „electric‟ and „magnetic‟ field distributions. The absence of
magnetic field at the source antipode is the most evident feature. Besides, a peak in the electric map
always corresponds to a relevant node in the magnetic map. This feature follows from the
conservation law of the total electromagnetic power which is equally distributed in the cavity, so
that a maximum of one field occurs at the minimum of the other field component.
2.12 ELF fields in the time domain
The time domain solution is constructed as the formal Fourier transform of the frequency
domain series (7) and (8). Spectral components of the fields have the poles in the complex
frequency plane, and so one can use the residual theorem for obtaining the ZHSR in the time
domain for both the fields (e.g., [Nickolaenko and Hayakawa 2002; Wait, 1970; Nickolaenko and
Hayakawa 2014]). Convergence of the ZHSR in the time domain is better than that in the frequency
domain, and still it might be accelerated by extracting the singularity. The results are of the
following form:
Copy right @ESES
IJEAR
9
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064



 itn  
Pn x 
Q

E t   E A  Re  A    B Q 1 Q0  1B   1  1 B B  1  B B  12 
exp   
n  B n  1  Aν  




n 1

 


1/ 2
-3/2
 it1  B  


H t   H A Im  1  x 2
1  2 x   2
exp 

A

ν




Here EA = MC/(2ha2) and
(11)
(12)
HA = MC/(2ha) are the electric and magnetic field amplitudes;
x = cos;  = exp(it/A) is the time factor; functions Q1 = (1  x)(1  2x + 2)3/2  1
and
Q1 = ln{[  x + (1  2x + 2)1/2]/(1  x)} were derived from the generating function of Legendre
polynomials Q0 = (1  2x + 2)1/2. Analytical time domain solutions (16) and (17) imply the linear
frequency dependence () = A  + B, so that the poles are positioned along the straight lines in the
complex frequency plane.
Each term in the series (11) contains a parameter, which decreases exponentially with time or
mode number n. This factor becomes equal to unity when t =0, so that the series (16) diverges at the
point  = 0 where Pn(cos)  1. Formulas (11) and (12) provide the finite function in the whole
Earth–ionosphere cavity for an arbitrary time t > 0, including the point where the lightning discharge
had been situated. This suggests a possibility of computing the waveform of an ELF pulsed signal
(Q–burst) returned to the point of parent stroke.
We depict the ELF pulsed signals in Figure. 4 computed with the use of Eqs. (11) and (12), in
which time from the stroke initiation is shown along the abscissa in milliseconds. Two left frames
demonstrate the E- (lower frame) and H-field (upper frame) as functions of time. Pulses originate
from the positive lightning stroke, i.e., when the electric current flows from the cloud to the ground,
so that the vertical electric pulsed field has a negative onset. The current moment of the source was
equal to 108 Am, and its spectral density was constant (the “white” source). Computations are based
on the heuristic linear frequency dependence of the propagation constant: (f) = (f – 2) / 6 – if / 100
(see below).
The left plots of Figure 4 show a set of pulsed waveforms corresponding to different SODs
varying with a step of 1 Mm. The curves were gradually elevated of the abscissa to facilitate a
comparison of pulses arriving from the distances of 5, 10, 15, and 20 Mm marked by color. One may
observe the direct pulse delay growing with the source distance and the antipodal pulse coming
closer and closer to the direct wave. At the antipodal 20 Mm distance, the magnetic field turns to
zero, while the electric pulse noticeably increases. Besides, the onset of electric field in the direct
and antipodal waves has the same sign while the horizontal magnetic field is a succession of „pulses
with alternating sign‟ (see [Nickolaenko and Hayakawa, 2002] for details). The upper right plot of
Copy right @ESES
IJEAR
10
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Figure 4 shows temporal variations of the round-the-world wave, which is the wave that returned to
the source point after the “round-the-world” trip. One may see that the propagation time is 159 ms,
so that the signal velocity of the pulse is 251.6 Mm/s. The ordinate shows that the pulse onset has an
amplitude of –2.81 m V/m, while the second half-wave is +4.3 m V/m. For comparison, the major
ticks on the ordinate of the lower left plots correspond to 100 mV/m, and this is the amplitude of
direct pulse at the distance of ~2.5Mm.
The lower right plot of Figure 4 depicts the distance dependence of the initial negative excursion
of the E(t) field together with a set of “standard” functions. In fact, this dependence might be
recognized in the sets of pulses collected in the left frames. Computations show how the electric
pulse amplitude reduces when the circular wave front extends from a point vertical source. The wave
front starts converging to the source antipode after passing the 10 Mm distance. This geometric
convergence of the wave is combined with its attenuation in the cavity. This is why the pulse
amplitude starts increasing from distances somewhat greater than 10 Mm. Figure 4 shows that the
electric pulse acquires the minimum amplitude at the source distance about 15.5 Mm. Since the
energy losses increase with signal frequency, the electromagnetic energy at higher frequencies
decreases faster, and the signal initially “white” acquires the “reddish” spectrum. This results in a
gradual widening of the pulse in the time domain. One may note that the Earth–Ionosphere cavity
acts like a low-pass filter with the lower cut-off approximately equal to 100 Hz. Geometric focusing
at the source antipode is seen as the electric field increases by approximately 10 dB from its
minimum level. Magnetic field is absent at the source antipode. The antipodal focusing of radio
waves may partially facilitate the detection of natural ELF transients arriving from great distances.
The wave front extension explains why the pulsed amplitude decreases faster than might be
expected from the sole exponential attenuation: |E| = exp(2) [/sin]½ [Nickolaenko and
Hayakawa, 2002 ; Nickolaenko and Hayakawa, 2014]. This dependence is shown in the right lower
frame by the green stars. It deviates from the amplitude of ELF transient because the monochromatic
signal does not experience the spectral modifications during the propagation.
We show only the first pair of pulses in the left lower frame of Figure 4: the direct and antipodal
waves. Here we also plotted a few characteristic functions. The red line shows the attenuation of a
plane wave with the rate of 3 dB/Mm. The blue, black and magenta lines illustrate the distance
dependence of 1/D, 1/D3/2 and 1/D2 correspondingly. One may observe that the distance variation of
ELF transient has its own distinctive outline.
Copy right @ESES
IJEAR
11
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
A pair of direct and antipodal pulses returns to the observer after each “round-the-world travel”
along the same great circle. This feature might seem odd, because the propagation takes place in the
spherical cavity where “all propagation paths are possible”. The following elementary consideration
demonstrates that this is not so. Let the vertical point source occupy the pole and emit a single short
pulse. The expanding circular wave front initially moves from the source being coincident with a
parallel. Then, it crosses the equator and starts converging to the source antipode. After focusing
there, the pulsed circular wave re-appears again and propagates from the source antipode to the
source point. It crosses the equator once again and merges at the source point. The process continues
until the pulse amplitude becomes negligible due to wave attenuation. Imagine that there is an
observer positioned at a meridian at some distance from the source. The observer will detect first the
circular wave front expanding from the source. In a time, the wave front arrives from the source
antipode or from the opposite direction. Thus, the observer detects pulses coming either from the
source or from the source antipode, and radio waves arrive in succession from two opposite directions
along the same great circle path.
Figure 4: Waveforms of the electric and magnetic pulses computed in the time domain
Copy right @ESES
IJEAR
12
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
2.2.
ISSN 2395 0064
ELF propagation
Aspects of ELF propagation have been reviewed in many books [Nickolaenko and Hayakawa, 2002;
Wait, 1970 ; Galejs, 1972 ; Bliokh, et al., 1980; Nickolaenko and Hayakawa, 2014; Jones, 1969;
Jones, 1970; Porrat and Fraser-Smith 2001; Polk, 1969; Ogawa et al., 1969; Volland,1995; Sentman,
1995; Harth 1982]. The ground surface might be considered a perfect conductor at ELF and the
ionospheric plasma is responsible for the particular phase velocity and wave attenuation in the
subionospheric radio propagation. The ELF radio waves are partially reflected over a relatively large
interval of altitudes. Owing to the low frequency, the „C-region‟ in the lower ionosphere formed by
heave ions and ion complexes produced by galactic cosmic rays plays a major role in determining
the losses in the cavity. The geomagnetic field turns the ionospheric plasma into an anisotropic
medium slightly modifying the phase velocity and attenuation and, particularly, the wave
polarization. ELF radio signals propagate with the velocity smaller than that of light, and the
attenuation rate varies from less than 0.1 dB/Mm at a few Hertz to about 60 dB/Mm at 2 kHz.
The propagation constant is the most important ingredient in the field computations. It was
deduced from the experimental data on SR or observations of ELF atmospherics by using the wave
impedance data. The simplest model describing the propagation constant as a function of frequency
is the linear model [Nickolaenko and Hayakawa 2002 ; Nickolaenko and Hayakawa 2014]. The
following linear model was found from SR cross-spectra measure at the Kharkov–Baikal-lake base
[Bliokh et.al .,1980]: (f) = (f  2)/6  if/100. This is a simple, but rather efficient model, and it was
used in the time domain solution of SR problem. We demonstrate below a high reciprocity between
observations and the time domain computations based on the linear model (f).
The (f) model suggested in [Ishaq and Jones, 1977] is considered to be the most accurate one. It
is described by the following relations:
 ( + 1) = (kaS)2,
(13)
where a is the Earth‟s radius, k is the wave number in free space, and S is the following complex
parameter:
S = c/V – i (5.59)  / f.
(14)
The real and imaginary parts of the S parameter are specified as follows:
Re[S] = c/V = 1.64 – 0.1759 ln(f) + 0.01791 ln2(f)
 = 0.063 f0.64.
Copy right @ESES
(15)
(16)
IJEAR
13
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
The propagation constant is found from Eq. (1), which is used in the computations of
electromagnetic fields in the Earth-ionosphere cavity.
This model is widely used in the SR computations, although a slightly different approximation
was derived for the wave phase velocity by [Huang et al., 1999] from fitting other data of SR
observations:
Re[S] = c/V = 1.725 – 0.275 lnf + 0.052 (lnf)2 – 0.0036 (lnf)3
(17)
In spite of visual deviations, Eq. (17) provides the (f) frequency dependence that departs from that
given by Eqs. (13)–(16) by less than 1% in the whole SR frequency band 4 – 40 Hz. Therefore, we
use in what follows the model [Ishaq and Jones, 1977] as the standard one.
2.3.
Approximate models of atmospheric conductivity
Exponential profile
Conductivity of the lower ionosphere is often modeled by an effective exponential profile Similar
models were used in the VLF band long ago [ Wait, 1970; Galejs, 1972], but these appeared to be
inappropriate for ELF where a special model was developed [Greifinger and Greifinger, 1978 ] for
monochromatic radio signals propagating in the flat Earth–ionosphere duct. It exploited the fact that
E and H fields penetrate into different heights of the ionosphere. The electric field starts to severely
attenuate after reaching the lower characteristic height h0 where the conduction and displacement
currents become equal to each other at the given frequency: E(h0) = 2  f 
(here  is the
permittivity of the vacuum). Magnetic field reaches greater altitudes, and it starts to rapidly decrease
from the "magnetic" characteristic altitude h1. Here the wavelength in the ionosphere becomes equal
to the local scale height of conductivity profile, i.e., the condition is held M(h1) = [4 012]1,
where 0 is the permeability of vacuum and 1 is the height scale in the vicinity of h1. The technique
implies the four parameters of the (h) conductivity profile; namely, two characteristic heights h0
and h1 and two scale heights 0 and 1 in the vicinity of characteristic altitudes for obtaining the
complex propagation parameter.
S = [h1/h0]½ [1 + i(0/h0 + 1/h1)/4]
(18)
The complex parameter S and ELF propagation constant at a given frequency depend on the four
parameters of the profile. Thus, one must find two functions of height relevant to a particular vertical
profile of atmospheric conductivity: the (h) conductivity itself and the scale height (h). The
Copy right @ESES
IJEAR
14
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
characteristic heights and scale heights are found from these profiles for the given frequency of radio
signal f. For example, when the frequency of monochromatic signal is f = 8 Hz, the "electric"
conductivity is E =2  8  = 4.444 10–10 S/m, and we can find the electric height h0 by using the
profile (h) and relevant scale height 0 from the (h) profile. Afterwards, we evaluate
M = 4.398104 S/m supposing that the scale height of the conductivity is 3 km. Thus we can find
the magnetic height h1 and the scale height 1. The search for the necessary parameters ends if 1 = 3
km. When 1  3 km, definite adjustments are necessary, so that any kind of iteration procedure
might be organized.
It is clear that such a procedure is inconvenient, and a simplified relation was suggested in
[Greifinger and Greifinger, 1978] for finding the magnetic height:
h1 = h0 20 ln(2k0).
(19)
This formula corresponds to the conductivity profile with a constant scale height, but it is also used
in the so-called two-scale models when the magnetic height is found by using Eq. (11), but the actual
magnetic scale height is used afterwards (see e.g. [Sentman, 1995].
Additional efforts were
necessary for extending this technique to the SR studies. The first, formal obstacle was connected
with the spherical geometry of SR problem, and it was not obvious beforehand that the results
obtained for the flat waveguide will remain valid in the spherical cavity. Fortunately it appeared that
formulas are correct in the spherical cavity, provided that the frequencies exceed a few Hertz
[Sentman, 1995; Nickolaenko and Rabinowicz, 1982; Nickolaenko and Rabinowicz, 1987; Sentman,
1990; Füllekrug, 2000]. The following relations were obtained for the SR frequencies and the Q–
factors:
f n  c 2a  nn  1h0 h1


Qn    0 h0   1 h1 1.
(20)
(21)
The second obstacle was the broadband nature of SR signal occupying the band up to 100 Hz and
even higher as seen in Figure. 3. Indeed, the above procedure becomes inconvenient when it is
necessary to find the variable characteristic heights and scale heights on a profile when the frequency
varies in a wide band.
Copy right @ESES
IJEAR
15
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
This problem was resolved by introducing the reference frequencies and relevant reference
heights. The lower and the upper characteristic heights are found from the following formulas at an
arbitrary frequency f:
h0 = Gh +  ln(f / Fg)
(22)
h1 = h0 2 ln(2k).
(23)
Here Gh is the reference height and Fg is the reference frequency. The only distinction between the
works using the exponential profiles lies in choosing the reference quantities. Various exponential
conductivity profiles were described in the literatures providing realistic models of ELF propagation
constant. Typical parameters of these profiles are: Gh 50 km and 0  3 km at the reference
frequency Fg = 8 Hz being equal to the first SR mode frequency. The magnetic characteristic height
is found from the equation h1 = h0 20 ln(2k0)  95 km. The scale height 1 might be taken equal
to 0 (the single scale model) or to some other value (two-scale model).
We must emphasize that the exponential profile model is just a convenient approximate tool for
computing the ELF propagation constant. No difficulties are met when the approach is applied to a
given function  (h) for obtaining the  (f) function. This result might be compared with the rigorous
full wave solution constructed for the same conductivity profile. This was done in [Jones and Knott,
1999] and it turned out that approximate formulas predict the real part of propagation constant
Re[ (f)] with the accuracy of 1–2%. But concerning the imaginary part Im[ (f)] or the wave
attenuation, the approximate value may deviate from the accurate one by 10–15%.
Knee model
A further development of approximate heuristic models was the “knee model” [Jones and Knott,.
1999 24]. It improved the relevance of obtained propagation constant to the standard dependence. In
distinction from the two-scale exponential model, the knee model operates with four frequencydependent parameters. Two of them are the complex functions of frequency interpreted as the
characteristic electric and magnetic heights, and the other two are the real functions of frequency and
we model the electric and magnetic height scales. The propagation constant  (f), which might be
used in the calculation of the fields, is computed using standard relations. The complex electric and
magnetic height hE and hM are found from the following equations at a given frequency f:
Copy right @ESES
IJEAR
16
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)


ISSN 2395 0064


hE  f   hKN   aln f f KN   ln 1   f f KN 2  a   b  2  i  a  2   a   b tan 1 f KN f 




hM  f   hm
  mln f f m*  i m  f  2

 m   m  bm 1 f  1 f *m
(24)
(25)
(26)
Here hKN = 55 km and fKN = 10 Hz are the reference height and the frequency of the knee. The
condition is satisfied at the knee height  (hKN) = KN = 2 fKN  where  = 8.8591012 F/m is the
permittivity of free space. Parameters a = 2.9 km and b = 8.3 km are the scale heights above and
below the knee height. The magnetic reference height and magnetic reference frequency are
hm*=96.5 km and fm*=8 Hz correspondingly. The scale height at magnetic reference frequency is
m*=4 km, and the frequency parameter of magnetic scale height is bm= 20km [Mushtak and
Williams , 2002]. Other parameters were also used [Pechony and Price, 2004].
The propagation constant (f) is obtained from the standard expression (13), which is equivalent
to the following equation:
(f) = [0.25+(kaS)2]1/2 – 0.5
(27)
The idea of the knee model is that at the frequency of the first SR mode, the characteristic
electric height hE lies below the knee altitude hKN where the scale height E is large. With the growth
in frequency, the electric height exceeds the knee altitude, and the scale height decreases. As a result
the quality factor Qn of SR is reduced at the fundamental mode and remains appropriate at higher
modes [Mushtak and Williams, 2002].
The propagation constant found in the knee model corresponds to the standard one, and it
became popular in SR computations, especially when the fields are obtained by using traditional
formulas (7–10).
However, parameters of the knee model were verbally referred to a certain conductivity profile
that never existed. At the same time the works appeared applying the direct modeling of radio
propagation by using FDTD or 3DTE (3 dimensional telegraph equation) techniques [Yang and
Pasko, 2005; Toledo-Redondo et al., 2013; Zhou et al., 2016]. The self-made conductivity profiles
(h) with a knee were used in direct computations and provided resonance spectra generally similar
to the observations. A closer inspection revealed unavoidable departures of these direct solutions
from “traditional‟ ones despite the both of them exploited the “same‟ knee model [Zhou et al., 2016].
It was demonstrated that the FDTD and the ZHSR give almost coincident results when applying such
Copy right @ESES
IJEAR
17
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
a self-made conductivity profiles (h) that follows from the knee model. In this case the propagation
constant was obtained for the profile with the help of rigorous full wave solution. Unfortunately, this
(f) dependence noticeably deviated from that found from the knee model by using Eqs. (24–27).
This result indicates that the heuristic knee model is rather good for approximating the standard (f)
dependence in the SR band. Simultaneously, it is unclear to what a conductivity profile (h) does
this model correspond, provided that such a profile exists at all. In other words, some other, realistic
conductivity profiles must be applied when using direct numerical modeling of SR. Any attempts to
derive a profile using the knee model result in unrealistic SR parameters. In what follows, we outline
a profile (h) corresponding simultaneously to observational SR and the global electric circuit data.
The propagation constant of ELF radio waves computed using formulas (25–28) of the heuristic
knee model is close to the standard model [Volland 1995; Sentman, 1995; Harth, 1982; Ishaq and
Jones, 1977 ]. Completely different data are obtained when computations imply the atmospheric
conductivity profile built with the help of characteristic heights and the scale heights of the same
knee model. Here, the effect is observed of the approximate physical interpretation of particular
elements of the heuristic model. So, this model should be used with caution when interpreting
observational data, because the deduced parameters of the profile will provide quite different results
when applied in rigorous computations. It would be desirable to use initially the realistic
conductivity profile, and such a profile is described in what follows.
Realistic profile
In the field computations and in the interpretation of experimental data, the knowledge is
redundant of the vertical profile of atmospheric conductivity. It is sufficient to use the regular
expressions for the electromagnetic fields incorporating the propagation constant, the current
moment of the field source, and the ionospheric effective height. Information on the atmospheric
conductivity (h) becomes obligatory when using the direct modeling such as FDTD or the 2DTE
(two dimensional telegraph equation) technique. The range of heights 30–100 km is crucial in the
ELF radio propagation, but it is inaccessible by modern remote sensing. Existing experimental data
on the air conductivity here are rare and have been usually obtained by the rocket probes. So, one
can find only a limited amount of altitude profiles of the air conductivity in the literature.
In interpretations and modeling the ELF data the classical (h) profile is often used [Cole and
Pierce, 1965]. This profile is shown in Figure 5 by the curve with dots, and the data are listed in
Table 1. The major drawback preventing its application in the direct modeling of SR spectra is the
Copy right @ESES
IJEAR
18
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
inaccurate propagation constant deviating from the standard model. The new profile [Nickolaenko et
al., 2016 ; Kudintseva et al., 2016] is shown in Figure 5 by a smooth line. This (h) dependence
provides a more realistic ELF propagation constant. Simultaneously, it does not seriously deviate
from the classical plot, and hence it does not seriously deviate from the traditional data.
The atmospheric conductivity versus height is shown in Figure. 5. The curve with points 1 is the
profile [Cole and Pierce, 1965] and the smooth curve 2 depicts the more realistic profile (h)
[Nickolaenko et al., 2016 ; Kudintseva et al., 2016]. Profile 2 has more pronounced alterations in the
50–60 km interval around the knee region. Deviations of two curves start at 30 km altitude, where
profile 2 is elevated over the classical plot.
Figure 5: Altitude profiles of air conductivity. Line 1 is the profile [Cole and Pierce, 1965 29] and line 2 is the
new profile corresponding to observations of SR and data on the global electric circuit.
The simplified conductivity profiles are widely used in direct methods of field computation.
These are typically the lg [(h)] plot incorporating the two straight lines forming a kink at the knee
altitude. The existence of the upper, “magnetic” characteristic height is often ignored, so that the
profiles applied in direct computations are mere two-scale exponential models. On the other hand, an
advantage of the direct FDTD, 2DTE, and 3D-TLM (3-dimensional transmission line model)
computational techniques is that these are able to incorporate arbitrary vertical profiles and even
combine them with some horizontal non-uniformities. Conductivity profile 2 shown in Figure 5
might be used in such computations. It was tested by using the full wave solution followed by the
traditional formal description of SR, and we show some of these data.
The ELF propagation constant  (f) is usually constructed on the assumption that the ionospheric
plasma is isotropic and horizontally homogeneous. The approach is justified by a great wavelength
of SR oscillations, because such long waves must be insensitive to “small details” of the cavity such
Copy right @ESES
IJEAR
19
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
as the localized or even global non-uniformities. A given conductivity profile corresponds to the
complex dielectric constant of the air separating the ground and ionosphere. This medium is
supposed to be horizontally stratified, and the electromagnetic problem is formulated for such a
configuration. The full wave solution of such a problem exploits waves propagating upward and
downward in each slab and the boundary conditions at each interface where the tangential field
components remain continuous [Bliokh et al., 1980; Jones and Knott, 1999 ; Madden and Thompson,
1965]. The full wave solution is the rigorous one, and it provides the system of 2N linear algebraic
equations for the amplitudes of transmitted and reflected waves in each of N layers. The thickness of
the layers is much smaller than the wavelength in the medium, and the problem might be
reformulated to nonlinear differential equation of the first order (Riccati equation) for the surface
impedance [Nickolaenko and Hayakawa, 2002; Hynninen and Galuk, 1972 ; Galuk et al. , 2015].
The latter is the ratio of the tangential components of E and H fields at the boundaries. When
reformulated in the Riccati equation, the surface impedance must satisfy the boundary conditions on
the ground and at an upper layer in the ionosphere.
Table 1 Logarithm of air conductivity (S/m) as a function of altitude above the ground surface
h km
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
lg()
.8
.66
.
.
.
.
.
.
.2
.
.
.
.
.
.3
.
.4
h km
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
lg()
.5
.
.
.2
.
.8
.
.
.
.

.
.
.
.
.
.4
h km
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
lg()
.
.4
.9
.
10.0
.
.2
.6
.3
.
.5
.
.
.
.
.
.6
h km
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
lg()
.
.
.4
.40
.
.3
.1
.2
.7
.
.
.5
.4
.1
.
.3
.50
h km
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
lg()
.5
.
.
.5
.
.5
.
.
.
.
.3
.
.
.
.
.9
.
h km
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
lg()
.9
.
.
.
.7
.
.
.
.
.9
.8
.40
.
.
.6
.4
.2
This method is regarded as the full wave solution, since it strictly accounts for all the fields
propagating in the stratified medium. The full wave solution provides the complex  (f) dependence
corresponding to a given profile (h). It is constructed numerically by using the iteration procedure
(see [Galuk et al., 2015] for details). Iterations in the full wave solution were performed until the
novel value of the surface impedance deviated from the previous one by less than 10–7.
Copy right @ESES
IJEAR
20
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
The classical full wave solution has a shortcoming that is revealed when applied in the SR
problem. The lower strata of atmosphere have very small conductivity, so that permittivity of air
stops varying with the altitude. The solution based on the application of waves in the slabs becomes
irresolvable in a sense that the algebraic linear system contains a set of the same equations
corresponding to lower layers, and its matrix becomes singular. This is why the full wave solutions
in the works [Jones and Knott, 1999; Madden and Thompson, 1965 ] started from approximately 30
km altitude while the lower slabs were somewhat thicker than the upper strata. This means
physically that ELF solutions thus obtained cannot be compared with the data based on
measurements of the fair weather electricity in the global electric circuit: these are relevant to
different altitude intervals.
This is not so when the full wave problem is formulated in terms of the surface impedance and its
solution is obtained by using the Riccati equation. One may include the atmospheric strata of equal
thickness ranging from the ground surface to the lower ionosphere edge. Physically this means that a
profile might be treated combining the information acquired from measurements both of global
electric circuit and the global electromagnetic resonance. This was done in [Nickolaenko et al., 2016;
Kudintseva et al., 2016]. Profile 2 in Fig. 5 presents this particular (h) dependence averaged over
the whole globe. It is clear that owing to the global nature of SR, its parameters predominantly
depend on the globally averaged properties of ionosphere. Data concerning the mean (h) profile and
its modifications to the ambient day and ambient night conditions might be found in [Nickolaenko et
al., 2016 ]
Frequency variations of the real and imaginary parts of the propagation constant are compared in
Figure 6 computed from the formulas of standard model (13–16) and from the full wave solution for
profiles 1 and 2 of Figure 5. One may observe that all models give very close values in the real part
of the propagation constant (i.e., the phase velocity of radio waves), and deviations do not exceed
one percent. However, the peak frequencies in the power spectra of SR are noticeably different (see
Figure 7). The reason is simple: a 1% deviation in the phase velocity is relevant to 0.2 Hz deviation
at the third SR mode of the 20 Hz peak frequency.
Departures from the standard dependence in the imaginary part or in the attenuation rate are more
distinct and they range from 1.5% to 25% for profile 1 and within the 6% interval for profile 2.
Plots in Figure 6 indicate that profile 2 provides the propagation constant close to the reference
model in the entire SR band. Therefore, this profile might be recommended for modeling the global
Copy right @ESES
IJEAR
21
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
electromagnetic resonance in the Earth–ionosphere cavity, especially when the direct methods of
field computations are used.
Figure 6: Dispersion curves  (f). Upper panels: frequency variations of the real part of propagation constant
in the vicinity of the first, second, and third SR modes, and lower panel: the imaginary part of propagation
constant.
Validity of the conductivity profile 2 might be also verified by comparing model wave attenuation
with the data of direct measurements of the man-made ELF radio signals. These monochromatic
radio signals were radiated by the US [Bannister, 1999] and the Soviet [Nickolaenko, 2008 ] ELF
transmitters. Data [Bannister, 1999] were based on the amplitude monitoring of the signal arriving
from the US Navy transmitter regarded as the Wisconsin Test Facility. The global network was used
in reception of the 76 Hz signal, and the deduced average attenuation rate was  =0.82 dB/Mm for
the ambient night and 1.33 dB/Mm km in the ambient day conditions. The average attenuation at this
frequency was equal to 1.08 dB/Mm, while the standard deviation due to seasonal variations was ±
25%. The imaginary part of the propagation constant at this frequency for the profile 2 is equal to
Im[ (f)]f=76 = 0.86 Napier/rad. By using the relation  = lg(e)Im()  1.346Im(), we obtain
that the relevant attenuation rate  (76 Hz) = 1.17 dB/Mm. This attenuation is very close to
observational data, which is certainly in favor of the model.
The imaginary part of the propagation constant was also published [Nickolaenko, 2008 ] for the
frequency of 82 Hz. The Im[ (f)] was equal to 0.92 Napier/rad, which corresponds to the attenuation
factor  (82 Hz) = 1.25 dB/Mm. This attenuation rate was inferred from the distance dependence of
the signal amplitude in the vertical electric field for radio waves emitted by the Kola peninsula
transmitter of the Soviet Navy. The model imaginary part of the propagation constant is
Copy right @ESES
IJEAR
22
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Im[ (f)]f=82 = 0.92, which is equal to the value measured experimentally. Thus, a comparison with
observations of the man-made ELF radio transmissions justifies the conductivity profile 2.
Figure 7: Spectra of SR (left) and altitude variations of vertical gradient of atmosphere potential (right)
The propagation constant is further applied in the field computations. To compare results
obtained with the profiles 1 and 2 and the reference model, we turn to the power spectra of the
vertical electric field component. To compensate the distance dependence of SR spectra, we used the
globally uniform spatial distribution of field sources. The power spectrum is described for such a
distribution by the following equation [Nickolaenko and Hayakawa, 2002; Bliokh, et al., 1980;
Nickolaenko and Hayakawa, 2014 ]:
E f  
2
   1

2 

n 0 n
2n  1
n  1    1 2
.
(28)
Here the “white” source is used, and the power spectrum Power E=|E(f)|2 is measured in arbitrary
units.
Three SR spectra are shown in the left frame of Figure 7. The black line with dots shows the
power spectrum obtained in the standard ELF radio propagation model, red curve with starts
corresponds to profile 1, and blue line presents data for the profile 2. These plots suggest that profile
1 generally overestimates the losses in the Earth–ionosphere cavity, so that the resonance pattern
becomes less pronounced and the peak frequencies and intensities are smaller than the observed
values. Profile 2 provides data closer to the standard spectrum, especially in the vicinity of the first
three SR modes that are usually used for inferring geophysical information from resonance data.
Copy right @ESES
IJEAR
23
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
The right panel of Figure 7 compares the height variations of the fair weather field. The black
curve corresponds to the conductivity profile derived from the global circuit data [Rycroft et al,
2000a; Rycroft et al., 2000b ], the red curve with stars shows the profile 1 data, and the blue curve
depicts results pertinent to profile 2. The altitude above the ground is shown on the ordinate in km,
and the abscissa depicts the decimal logarithm of vertical electric field in V/m. The strength of the
fair weather field was obtained from the Ohm‟s law: by multiplying the air resistance (h) = 1/ (h)
by the density of leakage current of the global circuit j = 1.5 pA/m2 [Rycroft et al., 2000b ]. Relevant
fair weather field on the ground level becomes equal to 99 V/m for profiles 1 and 2, but it is only 14
V/m for the profile [Cole and Pierce, 1965]. After integrating the field strength over altitude, we find
that the ionosphere-to-ground potential difference Vi is equal to 263 kV in the model of global
circuit, and 247 kV for profile 2, and only 61 kV for the profile in [Nickolaenko et al., 2016 ;
Kudintseva et al., 2016]. The latter value is exceptionally low. Thus one may conclude that profile 2
provides realistic results when used for the field computations in the frequency band ranging from 0
to 100 Hz at least.
3.
Natural ELF waves
3.1. Stationary ELF waves (SR)
Sources
Signals of global electromagnetic resonance are detected in the power spectra of vertical electric or
the horizontal magnetic field components. Sources of natural ELF radiation are the lightning strokes
occurring worldwide. All necessary details on the terrestrial lightning might be found in
[Nickolaenko and Hayakawa, 2002 ; Nickolaenko and Hayakawa, 2014; Porrat and Fraser-Smith,
2001; Polk, 1969; Ogawa et al., 1969; Volland, 1995; Sentman, 1995; Harth, 1982 ; Rycroft et al.,
2008]. Because of the low attenuation, the ELF pulse emitted by a lightning discharge is able to
multiply return to the observer after “circling the globe”. Natural ELF radio noise is a superposition
of pulses arriving from lightning strokes at a rate about 100 events per second.
SR is observed as a succession of peaks between 4 and 40 Hz in the power spectra of E and H
fields. The upper limit of SR observations is conditioned by the radiation of power supply lines (50
or 60 Hz), and we observe the geomagnetic pulsations and ionospheric Alfven resonance below this
band [Surkov and Hayakawa, 2014]. The major source of SR electromagnetic energy is radiation
from lightning strokes, and owing to low frequencies substantial contribution is made by continuing
currents of lightning. Only the strokes with large continuing currents are detected at ELF, which
Copy right @ESES
IJEAR
24
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
might pass unnoticed in the other frequency band. So, data on the global lightning distribution might
deviate when inferred from the SR and from the VLF band (a few tens of kHz frequency)
Individual discharges are random, and their parameters may vary substantially. Even nowadays,
finding the global distribution of lightning stokes is not a simple task (e.g. [Hutchins et al., 2012]).
Thunderstorms occur in many locations, and there are approximately 2,000 of them over the whole
Earth. The strokes tend to occur over the land rather than over the sea. Therefore, there exist three
global thunderstorm centers: the maritime continent in the South-East Asia, equatorial Africa, and
America. Maximum of the local lightning activity is observed in the afternoon hours, so that the
global thunderstorms move around the globe. Daily activity starts in Indonesia during morning hours
of universal time (UT), it “steps” over the Indian Ocean to Africa in the afternoon and moves to
America in the evening. Global thunderstorms cease late in the UT night when the Pacific Ocean
occupies the sunny side of the globe. Seasonal and annual variations are caused by motion of the Sun
in the sky being simultaneously connected with the local climate and weather.
The ELF waves are sensitive only to the „large scale‟ elements in the thunderstorm distribution.
At the first 8 Hz SR mode, the „insignificant details‟ like continents are not resolved. So the simplest
model of a single lightning center is successfully used at the first resonance mode. The single
equatorial global thunderstorm center is usually placed at the 17-18 hr local time, and it has the size
from 1 /4 to 1/3 of the Earth‟s circumference. Such a source follows the Sun during a day and
slightly drifts to the north or to the south in summer and winter.
About 2,000 thunderstorms are permanently active over the planet, each of them covering the
characteristic area of 1,000 km2 and maintaining the rate of global lightning of 50–100 strokes per
second. A thunderstorm cell provides the median current of 1 A to preserve the equilibrium in the
global electric circuit. A median lightning stroke transports a negative charge of about 20 C to the
ground, and a thunderstorm generates a median return stroke every 20 seconds. More detained
information on the global lightning activity might be found in the special literature (see [Polk, 1969;
Ogawa et a., 1969; Volland, 1995; Sentman, 1995; Rycroft et al ., 2008 ; Rycroft et al., . 2000b ;
Hutchins et al., 2012 Rakov and Uman, 2013] and references therein).
Instrumentation
Owing to the huge man-made interference, a SR observatory should be placed far away from the
industrial regions, transportation facilities, power supply lines, and populated area. A typical SR
observatory exploits a vertical electric antenna and a couple of orthogonal horizontal magnetic
Copy right @ESES
IJEAR
25
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
antennas. The detailed description of ELF antennas might be found in the monograph [Burrows,
1978]. The vertical electric antenna is an electrode (usually spherical) mounted on an insulator.
Specific geometric shape of the electrode is not important, and it is desirable that it has a large selfcapacity (20-30 pF at least) which might be easily calculated. The electrode is connected to the preamplifier with a typical input impedance of 109 Ohms, and the input capacitance of less than 100 pF.
Afterwards, the SR signal is fed to the receiver with a gain of 60 dB and a dynamic range of at least
120 dB.
The electric antenna must have the frequencies of mechanical vibrations beyond the working
frequency band. Usually this means that carrying construction is not very rigid, and the antenna
active electrode is rather massive. The effective height of a vertical electric antenna increases with
elevation of its active electrode regardless the conductivity of the carrying construction: the mast
might be made of an insulator or a metal. One may imply that the effective height of vertical electric
antenna is approximately equal to one-half of its elevation above the ground. The exact value should
be established by using the procedure of antenna calibration (see [Nickolaenko and Hayakawa, 2002
; Nickolaenko and Hayakawa, 2014 ] for details).
The ELF magnetic antennas are usually the induction coils with many turns winded on the
ferromagnetic core of high magnetic permeability. Antennas have the sensitivity angular pattern of
lemniscata (the 8–form), and one has to use a couple of orthogonal antennas to detect signals arriving
from any arbitrary direction.
The voltage induced in the magnetic antenna coil is the product of the time derivative of the outer
magnetic induction B and the effective area of antenna winding Seff. The design of magnetic antenna
must provide the greatest possible effective area, which depends on the geometry of the core (the
length and diameter), the material of the core, number of turns, and the type of winding. The latter
controls the average cross-section of a turn and the mutual capacitance of turns. This capacitance
combined with the antenna inductance causes undesired resonances in the magnetic antenna
sensitivity. The number of turns N usually reaches 100,000. The winding is split into ~10 sections set
at the cover of the ferromagnetic core. The core is usually ~1 m long and has a few centimeters in
diameter. It is made of a ferrite or the isolated permalloy leafs firmly placed in the plastic wrapping.
It is known that a ferromagnetic body “draws in” the outer magnetic field lines. The volume from
which the field is collected increases when the core becomes longer (its orientation is parallel to the
outer field). The following approximate formulas are valid for a majority of magnetic antennas. The
effective permeability of the core is:
Copy right @ESES
IJEAR
26
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
eff = (ml/mp)2 [ln(2ml/mp)  1]1
ISSN 2395 0064
(29)
where ml is the half-length of the core, and mp denotes the core diameter. The effective area might be
evaluated from the following.
Seff =  (ml)2 N /3
(30)
The physical meaning of the last equation is that the effective area of a magnetic antenna with a
ferromagnetic core is approximately equal to that of the air loop antenna having the diameter equal
to the semi-axis of the core ml. These formulas are approximate and correspond to the material
permeability of the core  >> 1. In practice this means that  > 1,000–2,000. These equations are
useful for preliminary estimates of a future antenna size, while the exact computations must be made
by using rigorous and more complicated equations [Burrows, 1978]. Eqs. (30) and (31) explain the
relative invariance of the magnetic antenna design: all antennas described in the literature have the
core 1–2 m long, the diameter of a few centimeters and N100,000 turns.
Another type of antennas is also used, and the most common is the horizontal wire antenna
grounded at its ends. Physically such an antenna is equivalent of a one-turn vertical loop antenna
(magnetic antenna) with the area of about its squared length.
The signal from the output of antenna pre-amplifier is usually fed through a twisted pair cable to
the SR receiver positioned in the laboratory room at ~100 m distance from the antennas. A typical
receiver is the low frequency amplifier with the gain about 60 dB and dynamic range not less that
120 dB. Its role is the signal conditioning for the Data Acquisition System (DAS), habitually placed
in a PC. This usually means that the receiver band-pass is formed by the standard high-pass and the
low- pass filters of 4 – 6 order, and the total amplification provides the RMS output signal amplitude
of about 1 V. The DAS exploits the analog digital converters (24-bit sampling is preferred) working
at the sampling rate a few times higher than the higher cut-off frequency of receiver. As a rule, the
SR receiver has the bandwidth 4 – 40 Hz, and the sampling frequency is about 200 Hz. The DAS is
often synchronized by 1 s time stamps from GPS, and this allows us to compare the records
autonomously performed at remote observatories. The digital record is stored in the PC memory for
further analysis and processing.
Copy right @ESES
IJEAR
27
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
The current processing and averaging of observational data are performed to reflect the major
results of SR monitoring, and these are displayed on the PC monitor. The first step in such a
processing is obtaining the power or amplitude spectra of the time domain records. This is done
using Fourier transform of the U(t) realization. To stabilize results of measurements, one usually
performs the data accumulation in the time domain, an „automatic‟ averaging takes place in this case
and the results become stable. An increase in the duration of the time domain record does not make
the spectral estimates stable. According to the localization principle of Fourier transform, an increase
in duration T of the time domain record simply raises the spectral resolution F = 1/T. Physically
this means that spectral densities become independent, provided that their frequencies are separated
by F or a greater amount.
This feature is seen in the plot of Figure 8 where the amplitude spectrum is shown of the SR
signal recorded by the receiver of 100 Hz band-pass with the sampling rate of 200 Hz at Tiksi
settlement (71.584N and 128.768E) [Ayurov and Bashkuev, 2016 ]. The amplitude spectrum in
Figure. 8a is based on the records of 470 minutes duration. We easily recognize the resonance modes
in the radio noise spectrum that looks like a noisy strip and an enormous peak relevant to the power
supply of 50 Hz.
The averaging procedure must be applied for the stabilization of spectra. This goal can be
achieved by applying the spectral windows convolving the “raw” spectrum and a special window
function. The windowing procedure averages the adjacent spectral densities and thus exploits the
principle of localization. Since the spectral densities are independent, averaging over the frequencies
becomes an equivalent to averaging over an ensemble of realizations with a concurrent reduction in
the frequency resolution. Averaging over the frequency in Figure 8a was performed with the narrow
spectral window of 16 frequencies. If one applies the wider spectral window, deviations at adjoining
frequencies will be reduced, and there will appear the SR pattern.
After applying such a procedure to the SR record, one obtains the hourly spectra shown in Figure
8b and 8c where SR modes are clearly seen. Here, the same receiver and the same sampling rate are
used. Duration of the record is 60 minutes now and the wide spectral window is used averaging
spectral data over 1024 points.
Copy right @ESES
IJEAR
28
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
A
b
C
Figure . 8 Fourier transforms of continuous ELF record at Tiksi (71.584N and 128.768E) [courtesy of
Ayurov and. Bashkuev]
There are many procedures of data processing that are based on similar, but other algorithms.
Most often, a record is divided into fragments of standard duration about 10 s, which provides the
0.1 Hz spectral resolution. Elementary spectra are computed by applying FFT to these fragments.
The power spectra thus found are averaged over 5 or 10 min intervals (30 or 60 spectra are averaged
in an hour). Such a processing reduces random fluctuations present in elementary spectra as these do
not repeat from sample to sample, while the SR pattern is continuously replicated in all the segments,
and so the resonance lines become rather smooth [Nickolaenko and Hayakawa, 2002 ; Bliokh, et al.,
1980 ;Nickolaenko, and Hayakawa, 2014].
Spectra of SR are observed in the power spectra of different field components. Sometimes, the
cross spectra are used. For example, the spectra <E(f)HX(f)*> and <E(f)HY(f)*> are the components
of the Umov–Poynting vector, where E(f) is the complex spectrum (Fourier transform) of vertical
electric field, HX(f)* and HY(f) are complex conjugate of complex spectra of orthogonal horizontal
magnetic field components (HX = HWE and HY = HSN), and the angular brackets denote averaging
over the ensemble of spectra of elementary realizations. The cross-spectra of magnetic fields
<HX(f)HY(f)*> and <HX(f)*HY(f)> detected at the same observatory are used in measurements of the
field polarization. The cross-spectra of records obtained at different sites are also used [Nickolaenko
and Hayakawa, 2002 ; Nickolaenko and Hayakawa, 2014].
Copy right @ESES
IJEAR
29
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
As a rule, the primary information in the spectral form is stored in the computer memory. The
secondary processing of records is performed with the spectra, and parameters of resonance modes
are deduced. Each mode is characterized by its amplitude, peak frequency and Q-factor or relative
width at the half-power level. The fine structure is always present in observed spectral data, and the
spectral parameters are deduced by using some kind of data smoothing. The fitting by a Lorentzian
curve is most popular, which is known in a mathematical analysis as a Cauchy curve or Cauchy–
Lorentz distribution function. It has a pole in the complex frequency plane so that y(f) = y0 + A/[(f –
fn)2+(½ fn/Q)2]1 where y0 is a constant, A is the resonance amplitude, fn is the resonance frequency,
and Q is the quality factor of resonance oscillations connected with the width of resonance line.
An alternative processing of spectra is the Rice approach [4] when the resonance line S(f) is
considered as the probability density function of the continuous frequency, so the relation is valid
f2
f2
f1
f1
fn2 = W2/W0 where W2   S  f  f 2df , W0   S  f df , and integration is performed in the vicinity of
spectral maximum. Usually, the intervals are used of 81.5, 141.5; 201.5 Hz, etc. Such a
processing is similar to the SR mode trackers [Nelson, 1967; Sátori et al., 1996].
4.
Some observational results
The above listed parameters of resonance are used in geophysical applications: median values and
the diurnal, seasonal, and inter-annual variations. The major cause of regular variations of SR
parameters is motion of the field sources relative to the observatory combined with a relatively low
quality factor of cavity. There is no problem to explain an origin of diurnal variations in the
resonance intensity by the motion of global thunderstorms around the Earth.
The motion is clearly seen in the daily variations of the median arrival angle of ELF radio waves
established using the orthogonal components of Umov–Poynting vector. Such records show both the
displacement of the lightning activity in respect to the observatory and the relevant increases in
intensity when the waves arrive from the global thunderstorm centers (see Figure 4.4 in
[Nickolaenko and Hayakawa, 2014]).
Figure 3 indicates that alterations in the SOD modify the observed intensity of resonance modes
due to spatial modal structure and the wave attenuation. These modulations impede monitoring of the
thunderstorm intensity itself. The easiest way to reduce the impact of modal structure is application
of resonance intensity integrated (summed) over three modes. Individual spatial distributions
Copy right @ESES
IJEAR
30
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
compensate each other, and an impact of SOD becomes substantially reduced in the integrated SR
intensity [Nickolaenko and Hayakawa, 2002 ; Nickolaenko and Hayakawa, 2014]. Application of
simultaneous records at widely separated observatories allows us to separate diurnal variations
occurring the universal and local time as seen in Figure. 9.
The signal processing technique applied was suggested in [Sentman and Fraser, 1991 ] and
improved in [Pechony and Price, 2006]. It exploits the simultaneous SR records at observatories
widely separated along longitude. Data presented in Figure 9 were obtained from the records of
Moshiri (44.4N, 142.2E), Lehta (64.4N, 34E), and West Greenwich (41.6N, 71.6W)
[Nickolaenko et al., 2011]. The output data are the coherent diurnal variations occurring in the
Universal time (UT), which reflect the pattern of global thunderstorm activity pertinent to different
months and years estimated from the SR records. The daily changes reach the factor of 2, and the
seasonal trend is of similar value with the maximum during boreal summer. The patterns for a month
tend to repeat year after year, which indicates the general reproduction of thunderstorm activity from
year to year.
Figure 9 demonstrates results of reconstruction of the source intensity from the global resonance
data for every month of the year except July. These data are missing due to strong interference for
local thunderstorms at the observatories: we did not have simultaneous records at all the three sites,
and so processing was impossible. Resonance intensity recorded at an observatory is a convenient
characteristic, which is often used in different applications, but it depends on two factors: the
intensity of the source itself and the SOD. The source intensity variations are present in all records
and occur coherently in the UT. Alterations caused by diurnal motion of thunderstorms around the
globe take place in the local time. Separating these two factors is a crucial task when working with
the SR intensity. It is clear that the Q-factor or relative width of resonance peak alters when the
height of a given peak changes against that of adjacent peaks. Thus, variations of SOD cause diurnal
modulations in the Q-factors. Simultaneously, the source motion in a cavity of relatively low qualityfactor modifies the position of resonance peaks over the frequency axis due to interaction of adjacent
modes. Diurnal frequency pattern is the simplest one at the first SR mode because the spatial
distribution of relevant fields is plain: it is proportional to the cosine or to the sine functions of
angular SOD. Correspondingly, the 1st mode frequency in the vertical electric field component has a
discontinuity at the nodal distance of 90 or 10 Mm. The discontinuity turns into a fast variation of
finite range df1 = fMAX – fMIN when the field sources occupy some area of width W [Wait, 1970 3].
One can estimate diurnal variations of the effective zone occupied by thunderstorms W by using the
Copy right @ESES
IJEAR
31
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
calibration curve W (df1) (see [Nickolaenko and Hayakawa, 2002; Nickolaenko and Hayakawa,
2014]). The first mode frequency in the horizontal magnetic field is proportional to the SOD, and
this dependence might be used to monitor diurnal, seasonal and inter-annual motion of global
thunderstorms [Nickolaenko et al., 2015].
Winter
Spring
Summer
Autumn
Figure 9: Diurnal/seasonal variations of global thunderstorm activity deduced from simultaneous SR records
at Moshiri, Lehta and West Greenwich. The data of July are missing due to the strong interference by local
thunderstorms at these stations.
Copy right @ESES
IJEAR
32
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
By applying the above properties of SR, the following estimates were obtained that agree with
other observations. The intensity of global thunderstorm activity varies by a factor of 2–3 during a
day and approximately by the same amount during a year, this latter variation being observed like a
trend. The two-three-fold inter-annual variations are also revealed by SR monitoring. The
characteristic size of a compact area occupied by lightning strokes varies from ¼ to 1/3 of Earth
circumference. In addition to the compact area, a half of global thunderstorms are uniformly
distributed in equatorial belt around the equator. Seasonal drift of thunderstorms is estimated by SR
data as 2–2.5 Mm.
4.1
SR as a global thermometer
The idea that SR intensity might be used as an indicator of current global thunderstorm activity was
spelled in [Williams, 1992], which demonstrated the similarity of the long-term alterations in
temperature anomaly and intensity of the first SR mode. Additional experimental evidence was
presented in [Sekiguchi, et al., 2006] showing a high correlation between SR intensity recorded at
the Moshiri observatory in Japan and the global soil temperature in different latitude intervals. The
following relation was obtained for the SR intensity I measured in dB relative to 1 pT2 and the
ground surface temperature in centigrade T measured in the latitude interval from 60S to +60N:
I [dB] =  13.64 + 0.648 T
(31)
The following relation is valid for the temperature deviations (anomaly) T from the median value
<T> = 16.96C in this interval:
I [dB] =  0.077 + 0.679 T
(32)
Relevant data are shown in the upper frame of Figure 10. The abscissa shows the months from
November 1998 to May 2002. The black curve with stars depicts the observed SR intensity in dB
relative to 1 pT2. The red smooth curve demonstrates the estimate for the resonance intensity
retrieved from the temperature observations in the [60S – 60N] latitude interval by using Eq. (32).
As one might observe, the correspondence of these two curves is rather good.
Copy right @ESES
IJEAR
33
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Figure 10: Seasonal variations of SR intensity and the ground temperature data
The lower plot of Figure 10 presents the situation with “inverse” modeling. Here, the ground
surface temperature was computed by using SR records (red line with dots) for the symmetric
latitudinal interval 20. Processing of ground surface temperature and ELF data collected at
Moshiri allowed obtaining the following relations [Sekiguchi, et al., 2006 ]:
PAN[dB] = 2cos(M/6) ,
(33)
where PAN is the annual component in variations of the resonance intensity at Moshiri, and M is the
month number (M=1 corresponds to January). The temperature anomaly T20 in the tropical belt
20S–20N is described by the heuristic equation:
T20[C] = 0.55cos(M/3)  0.3cos(M/6) = 0.55[10.5(PAN)2]  0.3{[10.25(PAN)2]1/2}
(34)
The red curve was reconstructed from the recorded annual variations of SR intensity by using Eq.
(34). The black curve with stars in the lower panel of Figure. 10, depicts the space-borne
observations of the ground surface temperature within the tropical belt. Again, outstanding
correspondence is present between the observed ground surface temperature anomaly and the
estimate deduced from the SR intensity. Unfortunately, the data available in [Sekiguchi et al., 2006 ],
Copy right @ESES
IJEAR
34
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
covered only 42 months of continuous observations. The comparison of longer records would be
desirable.
4.2
“Terminator effect” in SR records
The term “terminator effect” appeared when the long-term records of signal amplitude (intensity)
were arranged over the month–local time plane [Melnikov et al., 2004 ]. We show this plot in
Figure.11d, and the data were collected at Mitspe Ramon observatory (31 N and 35 E) in Israel.
The frame (d) depicts the seasonal–daily variations of the amplitude of the first SR mode recorded in
the vertical electric field component averaged over four years of observations (1999 – 2002). The
months are shown along the abscissa ranging from Jan. 01 to Dec. 31, and the local time is plotted on
the ordinate in hr. The color indicates the relative field amplitude, with higher levels being red.
The characteristic lens-like form of diurnal–seasonal variations was present in the records of three
SR modes recorded at many observatories, e.g. Nagycenk (47.6N and 16.7E). The observational
data showed the simultaneous increase in the electric and magnetic fields reaching 30%. The lenslike patterns were attributed to the influence of the day–night non-uniformity of the Earth–
ionosphere cavity because the abrupt increase/decrease in amplitude was found in the vicinity of
local sunrise and sunset. The day–night interface is called solar terminator; therefore the
phenomenon got the name of “terminator effect”. The above work did not support the conclusion by
computational data, while the published model data showed that the effect of ionospheric day–night
non-uniformity is about 5–10% at SR. The point is that since the SR is a global phenomenon, its
amplitude depends predominantly on the level of thunderstorm activity, and therefore the daily
patterns at different observatories must have the universal character. The day–night non-uniformity
provides a secondary effect.
The picture of lens–like variations made a great impression, and similar variations were found in
the records of many sites. However, none of experimental works discussed a possible formal
interpretation, and these ignore the model results [Rabinowicz, 1988; Pechony, 2007; Pechony et al.,
2007; Nickolaenko et al.; 2006a, Nickolaenko et al., 2006b]. The model data [Pechony, 2007;
Pechony et al., 2007] are shown in panels a – c of Figure. 11, relevant to the same Mitspe Ramon
observatory. Computations shown in this figure were made for the uniform Earth–ionosphere cavity
with the help of 2DTE.
Copy right @ESES
IJEAR
35
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Before discussing the data of Figure 11, we mention the major features relevant to the genuine
terminator effect.
1. The impact of ionospheric non-uniformity might be observed in the two distinct propagation
geometries: when the propagation path is positioned in the ambient day or in the ambient night
conditions. According to the wave diffraction concept, a propagation path is the source–observer
great circle arc surrounded by the first Fresnel zone of the field. One can speak about propagation in
the day (night) conditions when the first Fresnel zone is found within the day (night) hemisphere.
This condition is not held at the first SR frequency when the SOD is a few Mm because the first
Fresnel zone covers the whole globe. Hence, the ELF radio wave „does no resolve‟ the day and night
hemispheres, and it propagates in an average uniform cavity (we can call it the “dusk” or the “dawn”
resonator). The formal day or night propagation conditions are realized at the fourth SR mode and
above, which occurs at frequencies above those addressed here. We will briefly discuss in what
follows the impact of day–night non-uniformity on the artificial ELF radio signals.
2. The second feature is the frequency dependence of wave diffraction pattern: variations become
faster and sharper with an increase in the signal frequency. When turning to SR modes 1, 2, and 3,
we note that the relevant wavelength becomes shorter by the factor of 2.5. This means that variations
caused by the wave diffraction at the third mode must be faster and have the period 1.6 times shorter
that those at the first mode. In case when the patterns observed at different modes have the
remarkable similarity, the driving mechanism was not the wave diffractions, and most probably it
was the effect of variable source intensity.
3. The ionospheric height varies from ~60 to ~ 90 km from day to night, which means that median
height is 75 km and the day–night asymmetry is 15 km or 20%. This modest deviation in the cavity
height will cause a redistribution of the electric and magnetic field amplitude: the electric field will
be concentrated on the dayside and the magnetic field on the night side of the globe. However, the
following conservation law must be satisfied: the complete electromagnetic power remains a
constant in the unit volume of the cavity | 0E |2 + | 0H |2 = const. The physical idea is simple, if this
condition is not held, the power will flow from the places of high concentration to those of its lower
value until the equilibrium is established. This feature means that an increase observed in the electric
field at a point is combined by the equivalent decrease of the magnetic field at this point, provided
that alterations were caused by the non-uniform cavity height. In other words, since the electric field
concentrates on the dayside of the lower height, the magnetic field is reduced here. At the same time,
the magnetic field relocates to the higher side where the electric field is smaller. Therefore, the lensCopy right @ESES
IJEAR
36
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
like structure driven by the cavity height must have the opposite sign (an increase versus a decrease)
when observed in the electric and in the magnetic fields. Again, a simultaneous modulation of the
both field amplitudes must be attributed to the variable intensity of the source.
Analyses of observations and modeling indicate that the “terminator” effect is reflection of the
everyday development/cease in the global thunderstorm activity observed at the sites occupying
specific longitudes. The daily onset of the global lightning activity in the South-East Asia is
coincident with the sunrise at the European observatories, while its ending in America West coast is
close to the passage of evening terminator over the Europe.
Three source models were used in the computations shown in Figure 11. The upper left frame (a)
shows the results for the point source of constant amplitude. The source is positioned at the latitude
retarded from the noon point by 3 hr (at the local time of 15 hr). Its latitude varies with the season by
22.5: the source is an equatorial one during the equinoxes, and it drifts to the North and to the
South in accordance with the boreal summer and winter seasons. The lens-like pattern is clearly seen
in the model data of Figure 11a, and the amplitude maximum is observed when the source
approaches the observer (15 hr local time). The point source model has a serious drawback: the first
mode amplitude also increases when the source of constant intensity is positioned near the observer
antipode (the local time of 3 hr). Figure 11b presents a more realistic source model when global
thunderstorms are concentrated at the same point, but they become active only over the land. The
antipodal peak vanishes in this case because the observer antipode is found in the Pacific Ocean
where thunderstorms are rare.
Figure 11 Comparison of the model and experimental data “terminator effect” at the Mitspe Ramon
observatory
Copy right @ESES
IJEAR
37
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
The map of Figure 11c was obtained in the framework of the source model based on the OTD
(Optical Transient Detector) data averaged over the period of 1999 – 2002. As one may observe,
patterns similar to observations are obtained in the model based on the realistic source models in the
uniform cavity.
Similar model data were obtained by using ZHSR for the fields in the uniform cavity with the
sources distributed in accordance to the OTD data for a given month of observations (see
[Nickolaenko et al,. ; 2006a, Nickolaenko et al. , 2006b] for details). The upper map in Figure 12
shows the yearly dynamics of the field source intensity derived from the OTD data. SR intensity is
shown of the vertical electric field component in three maps were constructed for the European
observatories Mitspe Ramon (31N and 35E), Nagycenk (48N and 17E), and Lehta (64N and
34E). The observatory name and the coordinates are printed at every map. The month is shown on
the abscissa, and the UT in hr is shown on the ordinate. The range of UT variations was specially
extended to 27 hr to clearly show the upper edge of the lens-like outline (the evening time). The
color scale used for inking of all the maps is shown on the right. It is obvious that SR intensity
reflects diurnal/annual changes in the global thunderstorm activity in the uniform Earth–ionosphere
cavity. All maps have the similar shape: transition from low to higher intensity occurs in the
European “morning hours” regardless latitude and longitude of the site. The transition from dark to
bright in all maps is governed by the source activity, and it is present in dynamics of SR records.
Similarly to observations, data computed in the uniform mode also leave an impression that
variation of the SR amplitude is linked to the motion of the day-night interface over the
observatories. However, since the cavity was spherically uniform the pattern arises purely from the
behavior of the field source. Moreover, if we compute the maps for other globally distributed sites,
they will contain an abrupt variation in the SR intensity that occurs when the solar terminator passes
over Europe [Pechony et al., 2007 Nickolaenko et al., 2006b].
A noticeable increase of the duration of thunderstorm activity in Figure 12 during the boreal
summer is explained by the outline of continents in the Northern hemisphere. Indeed, with the
increase of northern latitude the western shore of the North America moves westward by
approximately 60 against the western edge of the South America. The Asian continent similarly
expands to the east, though to the minor extent. This is why the world thunderstorms start earlier and
end later in the summer months, so that the intense SR signal is recorded for a longer period at all
observatories. Thus, seasonal changes of high lightning activity are similar to the day duration in
Copy right @ESES
IJEAR
38
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Europe: the increase of Asian thunderstorm activity starts in-phase with the sunrise in Europe while
the decrease of the American activity occurs around the sunset in Europe.
An impact of the day-night interface on the SR oscillations was addressed in [Madden. and
Thompson, 1965] and [Bliokh, 1980]. The results of analysis were summarized in the monographs
[Nickolaenko and Hayakawa, 2002 ; Bliokh, 1980 ; Nickolaenko and Hayakawa, 2014] and in Ph. D
works [Rabinowicz, 1988 ; Pechony, 2007]. The terminator effect in the SR band cannot be
prominent because the pulsed wave front launched by a lightning stroke a few times passes through
the day and the night hemispheres prior to its complete attenuation. The kind of “averaging” takes
place as if the signal propagates in some effective uniform cavity. Depending on the smooth or the
sharp terminator model, the day–night non-uniformity is able to cause amplitude modulations from
5% (the smooth terminator) to 10% (the sharp step-like interface) (see [Nickolaenko and
Hayakawa, 2002; Nickolaenko and Hayakawa, 2014; Rabinowicz, 1988 ; Pechony, 2007; Pechony et
al ., 2007 ; Nickolaenko et al., 2006a; Nickolaenko et al., 2006b]. Some special experimental
techniques should be elaborated for detecting the tiny variations caused by terminator in the natural
radio noise having the 2-fold daily variations caused by the source motion and intensity.
Figure 12: Model lens-like seasonal–diurnal patterns computed for a set of observatories in the uniform
Earth–ionosphere cavity with the OTD data
Copy right @ESES
IJEAR
39
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
4.2
ISSN 2395 0064
Antarctic 10 years record
SR enables us to obtain characteristics of the lower ionosphere and dynamics of global
thunderstorms by using the data recorded at a single or a few observatories. Here, we apply here the
long-term SR record from the “Akademik Vernadsky” Ukrainian Antarctic station (65.25S and
64.25W) for demonstrating how the north-south seasonal drift of global thunderstorms might be
evaluated. A complete description and all results deduced from these records might be found in
[Sátori, 2003].
Two horizontal orthogonal components of magnetic fields are monitored by induction-coil
magnetometers in the frequency band 0.1–300 Hz. Antennas are oriented along the geographic
meridian (SN channel) and the parallel (WE channel). GPS synchronization was used. The
experimental data are processed as follows. Daily records from SN and WE channels are subdivided
onto 10-min intervals. The mean power and the cross spectra are obtained in these intervals with the
frequency resolution 0.1 Hz. The daily records contain 144 spectra.
The spectra undistorted by interference are selected and monthly averaged diurnal data sets are
derived each containing 144 spectra. These spectra are used to compute the monthly averaged
intensities, peak frequencies (for the SN and WE channels separately), and polarization
characteristics: the polarized intensity, ellipticity, position angle, degree of polarization. The modal
intensity and frequency are defined by using the above-mentioned Rice definition. Figure 13 surveys
the data corresponding to the first SR mode.
The upper frame in this figure depicts the long-term variation of the normalized intensity of the
first SR mode observed in the horizontal magnetic field I1 = I1WE + I1SN. The thick line in this panel
is the running average in the time window of 3 points (months) width. The middle frame in
Figure 13 presents variations of the first peak frequencies f1WE and f1SN observed in two orthogonal
magnetic field components. The line with crosses corresponds to the HSN field and the line with dots
shows the HWE component. To reduce rapid oscillations of peak frequencies, we also applied moving
average of the same width. Dashed lines depict the peak frequencies averaged over the year. The
lower frame illustrates the concurrent variations in the solar activity. The blue lines depict the
intensity of solar radiation at the 10.7 cm wavelength (the left ordinate), and the red lines show the
Wolf‟s number (the right ordinate). The dashed lines in the middle and lower frames present the
annual average. One may observe that solar activity reached the minimum in 2009 and the SR
parameters change in accordance with the solar cycle
Copy right @ESES
IJEAR
40
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
We use the recorded peak frequencies f1WE and f1SN for evaluating the long-term variations of the
effective SOD. The point source model was used for the world‟s thunderstorm activity. On the first
stage, the power spectra of magnetic field were computed at the Antarctic station by the source
positioned at a variable distance D. The model of the Earth-ionosphere cavity was used uniform
along the angular coordinates, and the ionosphere was isotropic. The knee model [Mushtak and
Williams, 2002] was applied for computing the ELF propagation constant, which was already
described in the previous section.
Figure 13 :Averaged intensities (top panel) and peak frequencies (middle panel) of first SR recorded at the
Ukrainian Antarctic Station “Akademik Vernadsky” (65.25 S and 64.25 W) from March 2002 to February
2012 with respect to solar activity indices (bottom panel) - intensity of solar radiation at the 10.7 cm
wavelength and the Wolf number
Results of computations are shown in Figure 14. The left frame shows the 2D intensity profile of
vertical electric and horizontal magnetic field components over the frequency–distance plane. The
abscissa shows the frequency in the vicinity of the first SR mode, and the ordinate depicts the SOD
in Mm. The dark inking shows the intensity of oscillations. These maps are in fact the fragments of
Figure 2 extended in the vicinity of the first SR mode. One may observe that the peak frequency of
electric field is discontinuous around the 10 Mm distance and that of magnetic field monotonously
increases with SOD. This allows us to evaluate the effective SOD from the observed value of
resonant frequency.
Copy right @ESES
IJEAR
41
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
We show the distance variations of the magnetic field peak frequency and intensity in Figure
14b. The horizontal axis depicts the SOD D in Mm. The first mode frequency is shown along the left
ordinate by the black line with dots. The line 1 with stars fits the distance dependence f1(D) by a
cubic polynomial. The normalized magnetic field intensity (the broken red line) is plotted along the
right ordinate in percents. Magenta line 3 shows variations of the first associated Legendre
polynomial I1(D) = [Pn1(cos)]2100% = sin( D/20)100%. The smooth curve 2 with diamonds
refers to the distance variations of the magnetic field intensity I1(D) by the polynomial of the fourth
order. The fitting polynomials were obtained by the least square technique and are expressed in the
following way:
f1(D) = a0 + a1D + a2D2 + a3D3
(35)
I1(D) = b0 + b1D + b2D2 + b3D3 + b4D4
(36)
Here D is the SOD in Mm, a0 = 6.71336, a1 = 0.181732, a2 =  0.008612582, a3 = 0.0001477123,
and b0 = 115.548, b1 = 30.40512, b2 = 7.1121217, b3 =  0.56481212, b4 = 0.0133658. By using Eq.
(7), one can correlate the first peak frequencies of 7.7, 7.8, 7.9, and 8.0 Hz with the source distances
of 8.3, 9.8, 12.2, and 16 Mm respectively.
When deducing variations of the effective source distance from the records of SR frequency, one
must use the inverse dependence D=D(f1):
D(f1) = d0 + d1f1 + d2f12 + d3f13
(37)
where d0 = 15003.1, d1 = 6033.82, d2 = 809.608, and d3 = 3.2566. Annual and interannual
variations of the effective distance from the source are shown in Figure 15. Here, abscissa depicts the
months from January 2002 to June 2012, and the evaluated effective source distance of shown along
the ordinate. The violet line with crosses was obtained from the f1SN frequency and the green line
with dots corresponds to f1WE frequency. The dashed lines depict the annually averaged distances
found from two orthogonal field components. These data are also listed in Table 2.
Variations of effective source distance were derived for the idealized point source model, and the
observed frequency variations are attributed exclusively to the seasonal and inter-annual drift of
global thunderstorms. In spite of such strong assumptions, we obtain a reasonable result. For
instance, seasonal northward drift is clearly visible in Figure 15 during the boreal summer: the global
thunderstorm centers move northward in the middle of a year, and we observe that the source
Copy right @ESES
IJEAR
42
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
distance D from the Antarctic observer increases in both the field components. The range of annual
drift is estimated as ~4 Mm from the SN field component (sensitive to African thunderstorms). Such
a displacement reaching 2 Mm or even higher corresponds to 18, which is quite reasonable.
Variations derived from the WE component (American and Asian thunderstorms) are less
pronounced: about 1.5 Mm or 13.
A
B
Figure 14 (a) Resonance intensity over the frequency–distance plane in the vicinity of the first SR mode; (b)
distance dependence of the first mode frequency and intensity: curve 1 – the I(D) = sin( D/20) dependence;
curve 2 – peak frequency; curve 3 – intensity.
Plots in Figure 15 show the reasonable range annual variations and the inter-annual, which might
be explained by alterations in the solar activity, but both estimates exceed the value of 10 Mm. The
large median source distances deviate from the expected values. There might be the following causes
of this discrepancy:
Copy right @ESES
IJEAR
43
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
(i)
ISSN 2395 0064
Different ways of evaluating the peak frequency. The values might deviate from those
obtained by fitting the resonance peak in the observed spectra.
(ii)
Deviations might arise from an irrelevant ionospheric model that underestimates the peak
frequency. Validity of such an assumption seems doubtful, since the knee model was specially
developed for the most exact description of the experimental spectra, although recorded in the
Northern hemisphere.
(iii) Application of the point source model, which ignores the spatial distribution of global
thunderstorms. The success of such a simple model is explained by closeness of the observed
peak frequencies to the interval where the f1(D) function is a linear one. In case of distributed
thunderstorms, the distance estimated corresponds to the centroid of zone covered by lightning
strokes.
The qualitative and quantitative agreement exists among the data on the seasonal drift of global
thunderstorms obtained with other ELF experiments (see e.g. [Sátori, 2003; Sátori et al., 2003 ;
Sátori
et al., 2005; Sátori et al., 2006 Sátori et al., 2009; Shvets et al., 2010; Shvets and
Hayakawa, 2011; Hayakawa et al., 2011 ]). Still, further modeling and relevant calibrations are
necessary to improve the accuracy of the source parameters evaluated by using SR records at
different sites. Direct comparison with the concurrent compatible records in the Northern and
Southern hemispheres is strongly desirable.
Figure 15 Annual and inter-annual variations of SOD deduced from the records at the Antarctic station
Copy right @ESES
IJEAR
44
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Table 2
Annually averaged peak frequencies of the first SR mode and relevant source distances
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
<f1> [Hz] in different field
components
HWE
HSN
8.016
7.966
8.015
7.971
7.999
7.958
7.984
7.958
7.969
7.940
7.958
7.930
7.943
7.927
7.931
7.916
7.951
7.941
7.967
7.953
SOD, Mm
HWE
16.62
16.56
15.52
15.29
14.7
14.3
13.76
13.35
14.05
14.63
HSN
14.6
14.77
14.3
13.66
57.43
13.32
13.2
12.87
13.67
14.13
Antarctic long-term records were exploited in [Sátori, 2003] for evaluating feasible inter-annual
temperature changes by using the following connection between the temperature anomaly of tropical
land and the intensity of SR [Sekiguchi et al., 2006]:
T [0C] = 0.0177 + 0.4845I [dB]
(38)
SR intensity in Antarctica has decreased by ~3.2 dB at the first resonance mode during the time
interval of 2003-2009 of reduced solar activity (see Figure. 13). Such a reduction results in the 1.6°C
decrease of the tropical land temperature. This trend has no link to the global warming or cooling.
This is a regular change in the soil temperature of the tropics related to the 11-year cycle of the solar
activity, which was not yet mentioned in the literature. The alteration evaluated is related to the
ground temperature, while 80% of tropics are covered by the ocean having the stable temperature, so
that the globally averaged estimate is reduced to ~0.3°C. This value is quite detectable, and it is
desirable to compare the available temperature and the SR data.
4.3
Disturbances associated with space weather
Solar proton events (SPEs)
Monitoring of SR provides interesting results connected with the abrupt changes in the Earth
environment, in the space weather. The effects are observed as quick abrupt changes in the records
superimposed on the regular diurnal pattern. These transient variations are usually caused by
modifications of the global ionosphere. In this sense, monitoring of peak frequencies has the obvious
advantage: an abrupt alteration in the peak frequency might be caused either by changes in the
Copy right @ESES
IJEAR
45
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
ionosphere or by the sudden displacement of global thunderstorms. The second modification is
doubtful.
Ionospheric disturbances at different altitudes may provide the opposite reaction of SR
frequency, which was noted in the pioneering study [Madden and Thompson, 1965] and multiply
confirmed afterwards (see e.g. [Yang and Pasko, 2005 ; Zhou, et al ., 2016] and references therein).
Two types of disturbances exist in the Earth–Ionosphere cavity, and both of them have the
cylindrical symmetry. One of them is the Sudden Ionospheric Disturbance (SID) and the other is the
Polar Cap absorption (PCA). An SID event is the abrupt enhancements in the dayside lower
ionosphere caused by the solar X - flares. The plasma modifications belong to the day–night type and
therefore they produce only minor changes in the peak frequencies hardly visible in the records.
Disturbances of the PCA type are caused by the charged particles arriving at the Earth from the
Sun. These are deflected to the poles by the geomagnetic field and precipitate in the polar and subpolar regions. The polar ionospheric D region goes down at both the poles, and a PCA is observed as
a radio black-out on the sub-polar propagation paths. The PCA provides a global ionospheric
modification that is symmetric in respect to the geomagnetic equator. This feature enhances its
impact on the SR frequencies.
Figure 16: Abrupt changes in the peak frequencies caused by the Bastille Day SPE
A solar flare starts with radiation of X-rays and gamma quanta, which overcome the 150 million
km distance in ~500 s or 8 minutes. The electromagnetic radiation is followed by the relativistic
Copy right @ESES
IJEAR
46
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
solar protons (SPE) or electrons (SEE (solar electron event)) which arrive with the delay of a second.
The proton energy may exceed the 10–100 MeV level, and that of the electrons exceeds a few MeV.
Abrupt initial changes are followed by modifications in the solar wind flux causing the geomagnetic
storms starting in a couple of days after solar flare initiation. Depending on the ionosphere
modification height, the particular PCA and SID events may appear in the SR records in different
ways, though both of them are caused by an increase in the lower ionosphere conductivity driven by
the solar activity. We describe the impact of SPE on SR record first and afterwards describe the
special processing technique that reveals the reaction of SR frequencies to the solar X- ray flares.
Figure 16 illustrates the effect of a powerful SPE the occurred around the UT noon on July14,
2000. The event is regarded in the literature as Bastille Day SPE. The abscissa shows the time in
days of July, and the observed SR peak frequencies are shown on the ordinates.
Two upper panels in Figure 15 show an abrupt decrease of the first and the second SR mode
frequencies recorded at Lehta observatory (64N and 34E). The vertical red arrow denotes the SPE
onset. One may note the rapid decrease of the peak frequencies of 0.4 Hz at the first, and 0.8 Hz at
the second modes. Two lower frames of Figure 16 show the records from the Karimshino
observatory (53N and 158E). Changes are depicted of the peak frequencies of the horizontal
magnetic field components H and D. The H–component lies in the plane of magnetic meridian, so
that it is close to the geographic HSN, and D – component is perpendicular to this plane, it is close to
the HEW component. Each plot is shown by thin and thick lines. Thin lines show the original data,
which indicate the power spectra accumulated every 10 min. Thick lines are the moving average (1
hr window) of quickly varying raw data.
Development of a classical PCA is clearly seen in Figure 16. The simultaneous global decrease of
SR frequencies is apparent, and changes start at arrival of solar protons to the Earth. The
simultaneous abrupt decrease in different field components recoded at widely separated field sites
indicates the global modification of lower ionosphere [Roldugin et al., 1999; Roldugin et al., ;
2001; Roldugin et al., ; 2003 Nickolaenko et al., 2008].
A close examination of Figure 16 reveals a short increase of peak frequency at the onset of the
SPE [Roldugin et al., , 1999; Roldugin et al., 2001; Roldugin et al., ; 2003 ; Nickolaenkoet al., 2008].
It lasts for 5–10 min followed by a reduction during 1 hr or longer. Such a development indicates
that initially the ionospheric conductivity grew around the 100 km altitude causing an increase in the
peak frequency [Bliokh, 1980]. Additional ionization penetrates to lower altitudes with time, and the
conductivity finally increases around the electric characteristic height causing the reduction in the
Copy right @ESES
IJEAR
47
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
peak frequency. Plasma modification covers the distance from 100 to 60 km in about 10 minutes,
hence, its vertical velocity is 4 km/min or ~70 m/s. Such a slow expansion indicates the diffusion
mechanism rather than the particle precipitation.
A comparison of record and the model data [Bliokh, 1980; Nickolaenko and Hayakawa, 2014;
Nickolaenko et al., 2008] allowed us to conclude that the Bastille Day SPE has caused approximately
50% modification of the ionosphere above the poles. Since the observed peak frequency depends on
the ratio Z / kh, where Z is the effective surface impedance of the lower ionosphere, h is its effective
height, and k is the free space wavenumber, the observed modifications could be attributed either to
the 50 % increase of effective surface impedance of the polar ionosphere or to the 50% reduction of
ionospheric effective height over the poles, or to the both of them simultaneously being 25% each.
4.4
Solar X-ray flares. Novel technique
The effect of SPEs on SR records is noticeable; therefore it is commonly addressed in many
publications. Impact of the X-ray flares is less pronounced, and a problem exists of its identification
at the background of stochastic variations relevant to the random nature of ELF field sources [Shvets
et al., 2016a ; Shvets et al., 2016b]. No convincing data were presented in the literature
demonstrating the influence of a solar X-ray flare on the SR frequencies (e.g. [Sátoriet al ., 2016]).
We introduce below a special technique that reveals a simultaneous sharp variation in all resonance
mode frequencies, which are usually hidden in the customary records. The technique implies the
weighted average frequency of SR, which is formally based on the mutual relation
fn = 5.7 [n(n+1)]1/2 of the modal frequencies mentioned when discussing Figure 2.
Impact of a X-flare on the SR pattern is readily explained physically. The hard solar
electromagnetic radiation sharply increases the plasma density on the dayside of the globe, so that
the boundary of the lower ionosphere goes down here. The night ionosphere is hidden in the shade,
and it remains undisturbed. Reaction of ELF waves to X-ray flare is caused by the reduction of the
globally averaged (effective) ionospheric height. Modeling shows (see the next-subsection) that such
a reduction causes a simultaneous drop of all eigen-values, and the effect is observed as a
modification of the whole SR pattern.
We present the SR records at the Karymshino observatory (52.9°N and 158.25°E) where three
orthogonal components of the magnetic field H, D, Z are recorded with the sampling frequency of
100 Hz. The elementary power spectra of these fields were computed by the Fourier-transform of the
fragments 3,000 points long. The 9-point moving average was used, and the results were averaged
Copy right @ESES
IJEAR
48
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
over the 5-minute intervals. The resonance parameters were obtained by fitting these spectra by a
sum of four Lorentzian curves (the first four SR modes).
Figure 17 shows data recorded during the X2.1 solar flare on March 11, 2015. The abscissa
depicts the universal time in hours. The thin fractured curves show the modal frequencies with the 5
min time step. Here, the regular daily variations of modal frequencies were found by averaging
records over 5 successive days (± 2 days relative the date of the event) and these were subtracted
from the observed daily pattern. To facilitate comparison, we separated individual modal plots
vertically by 0.5 Hz. The dotted line depicts the flux density of X-rays (I) shown on the logarithmic
scale on the right ordinate. The X-ray data were obtained by the GOES-15 satellite.
Figure17:Temporal variations of four modal frequencies during the solar X-ray flare on March 11, 2015 with
the mean daily runs subtracted. Variations of the X-ray flux power are shown by a dashed curve.
An optimist would note an abrupt increase in all peak frequencies at the onset of the flare.
However, the reciprocity to the solar flare is hardly convincing owing to many false abrupt variations
that mask the only one excursion relevant to the flare. Thus, one cannot register credible reaction of
SR frequencies to the X-flare.
Simultaneously, Figure 17 demonstrates that random variations in individual peak frequencies
deviate from each other. In contrast, the reaction of peak frequencies to the abrupt change of global
ionosphere must be similar and simultaneous at all resonance modes. We will reduce the random
fluctuations and reveal the coherent reaction by special averaging of peak frequencies. For this
purpose, we introduce the weighted resonance frequencies f(n)=fPEAK[n(n+1)/2]1/2, which are
approximately equal to 5.7 Hz (see Figure 2). Averaging of the weighted frequencies over a few
Copy right @ESES
IJEAR
49
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
modes must reduce the fluctuations. Simultaneously, the weighted averaged frequencies remain
sensitive to the global properties of the lower ionosphere.
Figure 18 shows excursions of weighted average frequency during the same X2.1 event of March
11, 2015. Plots in this figure demonstrate now a close coincidence of the main increase in the X-ray
flux and the reaction in weighted average frequencies. The similarity becomes apparent owing to a
substantial decrease in frequency fluctuations. One can even observe a detailed agreement of two
records: the reaction became visible on some secondary increases in the X-ray event.
Figure 18: Temporal variation of weighted average frequency during the solar flare of March 11, 2015
To demonstrate the reaction of SR modes to the global modification of lower ionosphere, we
place the black wedge and the triangle in the plot. The first one indicates the moment of sunset at the
observatory, and the second one corresponds to the local sunrise. Clearly, reaction of SR frequencies
was recorded in the ambient night conditions while the X-rays modified the dayside ionosphere.
Such a reaction would be impossible at VLF. Thus, we conclude that an application of weighted
average frequencies of SR reveals the reaction to the SID events.
4.4
Powerful gamma-ray flare and SR background spectrum
We briefly discuss the impact of the powerful gamma ray flare from SGR 1806-20 on SR. The
flare occurred around 21 hr 30 min and 26.5 s UT on December 27, 2004. The hard X- and gammarays arrived at the dayside of the Earth covering almost the whole hemisphere. Radiation came from
a soft gamma repeater (SGR) neutron star SGR 1900+14 positioned 30–40 thousand light years away.
The peak flux exceeded the most intense solar flares by approximately 5 orders of magnitude, and it
was 100 times greater than the SGR 1900 + 14 gamma flare of 1998. The ionospheric disturbance
was centered above the Pacific Ocean, at the geographic coordinates, 146.2W and 20.4S, which
Copy right @ESES
IJEAR
50
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
was only 450 km away from the center of the dayside hemisphere, and it covered the whole
hemisphere. The gamma flare lowered the dayside ionosphere by ~20 km, and this abrupt reduction
lasted less than 0.02 s. The temporal change of the height was fitted by the function:
dH = 19 (9.6441 t)0.1501 (see [Tanaka et al,. 2011; Nickolaenko, et al., 2012]) where t is the time
since the gamma ray arrival. The ionospheric height was varying with a step of 10 s in the model
computations. Information in more detail and complete reference might be found in [Nickolaenko
and Hayakawa, 2014].
We describe in what follows the successful detection of changes in the SR spectra during the
gamma flare on December 27, 2004. The record was performed at the Moshiri observatory.
Experimental material is compared with the model computations, and it was crucial to have the
model results prior to inspecting the data: modeling has shown what we should look for in the
record.
The heuristic knee ionosphere model was used [Mushtak and Williams, 2002 ]. To incorporate
the ionosphere day–night non-uniformity into modeling, the partially uniform knee (PUK) model
was used [Pechony and Price, 2004 ]. The computations suggested that the knee altitude was reduced
by gamma rays in accordance with the above formula. Thus, the electric height of the ionosphere
goes down, while the magnetic height remains undisturbed. Such a modification of the sole electric
height is discussed in [Inan et al., 2007] based on the VLF monitoring at Antarctica “Palmer” station.
The 20 km ionosphere reduction over the dayside of the globe results in the average height decrease
of 10 km.
By using 2DTE modeling, the results were compared of the 10 km reduction in the uniform
cavity and the 20 km reduction in the cavity with the day-night non-uniformity [Pechony and Price,
2004]. These two ionosphere models incorporated the globally uniform thunderstorm distribution
and the lightning distribution following from the OTD data. A comparison of the model results
showed that the day-night non-uniformity provides a secondary effect in the general modification by
the galactic gamma rays in comparison with the reduction of globally averaged ionosphere.
Figure 19 shows the impact of gamma rays computed in the models the non-uniform Earth–
ionosphere cavity (upper plot) and of the uniform cavity (lower plot). Individual spectra were
computed for the vertical electric field by using the 2DTE with the PUK day-night model [Pechony
and Price, 2004 25]. The OTD source distribution was used in both cases. Plots of Figure 19 have the
same feature: in all models, the dynamic spectra contain the characteristic abrupt drop in the
resonance pattern, which is coincident with the arrival of gamma rays.
Copy right @ESES
IJEAR
51
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Figure 19: Changes in the SR sonograms at the moment of gamma flare computed for the Earth–ionosphere
cavity with the day–night interface (upper frame) and the uniform resonator (lower frame)
Figure 20: Sharp amplitude variations during gamma flare computed for the uniform and non-uniform cavity
models with the uniform and OTD spatial distributions of global thunderstorms
Abrupt amplitude variations are given in Figure 20 obtained with the 2DTE for the amplitude of
the first SR mode. We plot here modulations for the uniform and the non-uniform Earth–ionosphere
cavities. The observer is placed at the “Palmer” Antarctic station, where the impact of gamma flare
Copy right @ESES
IJEAR
52
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
on the lower ionosphere was detected in VLF transmissions [Inan et al., 2007]. Again, both uniform
and the OTD spatial distributions of lightning strokes were used combined with the uniform and daynight models of the Earth–ionosphere cavity.
The inset in Figure 20 shows the amplitude reduction with the higher temporal resolution, t  [0,
10 s] where t = 0 is the gamma flare onset. One may conclude by comparing the black line (OTD
lightning distribution in the non-uniform cavity) with the blue one (OTD distribution in the uniform
cavity) that impact of the ionosphere day-night asymmetry plays the secondary role. The general
behavior remains the same: an abrupt reduction reaching 30%, which is followed by gradual
recovery. The highest deviation is pertinent to the uniform distribution of lightning strokes. Thus,
computations predict essentially the same modification for all models: gamma rays abruptly and
simultaneously reduce all resonance parameters at all resonance modes. Hence, the effect must look
as a discontinuity in the observed SR sonograms around the arrival time of gamma flare.
We use the ELF monitoring at Moshiri. Waveforms of vertical electric, the west-east (WE), and
the south-north (SN) horizontal magnetic field components are recorded here with the sampling
frequency of 4 kHz [Tanaka et al., 2011]. A fragment was chosen of the ELF record in the vicinity of
the gamma flare and the dynamic spectra of HSN and HWE components were found as seen in Figure
21. Time is shown on the abscissa in seconds, and the zero time corresponds to the moment of
gamma ray arrival. The ordinate depicts the frequency with the step of 2 Hz. We present two survey
sonograms in this figure.
Apparently, the data processing must satisfy two contradicting conditions. The model
computations predict the abrupt shift of resonance pattern by about 1 Hz, which lasts less than 10
seconds. To detect such a shift, the frequency resolution must be 0.1 Hz at least, so that segments of
10 s duration should be used in the FFT procedure. This is rather a long duration, because the
ionospheric disturbance exists for less than 10 s. To obtain the necessary time and frequency
resolution, we had used the time segments 10 s long, but the individual spectrograms were obtained
with the 1 s step. In particular for the moment t = 0, we process the segment ranging from 5 to +5 s;
for the time t = 1, we use the segment from 4 to +6 s, etc. These “central” times are shown on the
abscissa of Figure 21.
Copy right @ESES
IJEAR
53
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Figure 21: Comparison of observed data with the model dynamic spectra in the vicinity of gamma flare
The sonogram in Figure 21 visibly demonstrates the above-mentioned instability of experimental
SR spectra associated with the stochastic nature of lightning strokes. The commonly known smooth
resonance lines appear after averaging of elementary spectra acquired in the 5–10 min interval (from
30 to 60 elementary spectra). Therefore, one hardly can expect a regular resonance pattern in Figure
21: individual spectra are spontaneous in both the field components and this is the major obstacle in
revealing the impact of gamma rays.
To disclose the global ionosphere modification, we combine in Figure 21, the experimental
sonograms with the expected (model) modification shown by the thick black lines corresponding to
the constant levels in the model sonogram of Figure 19. The sonogram covers the time interval 30 s
around the gamma flare, and one may observe the similarity of expected and observed modulations
in both the panels of Figure 21. Time of sudden reduction in the SR record fits the gamma ray
arrival. Even the tendency toward the recovery might be perceived in Figure 21.
In contrast to the SID from solar X-ray flares, the ionosphere modification by the galactic gamma
rays was much more intense, but of short duration. This swiftness is a serious obstacle in the
experiment. Only direct comparison of computational and experimental sonograms revealed the fast
simultaneous modification in the SR pattern caused by the powerful gamma ray flare of December
27, 2004. We can argue that conventional signal processing will miss the modulation. A special FFT
procedure was necessary combining the 1 s time step with the 10 s segments.
Storms in the geospace result in different types of global ionosphere modifications. When a
disturbance is driven by the hard electromagnetic radiation, its effect enhances the day-night nonCopy right @ESES
IJEAR
54
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
uniformity of the ionosphere (SID). Modification concentrates at the poles in both hemispheres when
ionization is produced by the corpuscular radiation deflected by the geomagnetic field (PCA). Of
course, this is a simplified picture, e.g., we do not mention the Earth‟s magnetosphere etc.. The
symmetric polar disturbances are more efficient for the SR parameters than the asymmetric day-night
modifications [Galejs, 1972; Bliokh et al., 1980; Nickolaenko and Hayakawa, 2014]. This is why the
SPEs are easily seen in the SR records, while the solar X-flares demand a special processing. An
event of the giant gamma flare on December 27, 2004 has caused the strong, but very short
modification that was barely detected in the spectra of global electromagnetic resonance.
Inverse problem: the rigorous approach
SR data are usually collected for their further application in environmental studies, i.e., in solving
the inverse electromagnetic problems. One can mention two kinds of these problems: those devoted
to the global thunderstorms and those directed to studies of lower ionosphere. We briefly illustrate
the first class of application.
The exhausting search is applied in the solutions aiming at the properties of lower ionosphere.
For example, we found the simultaneous modification of resonance frequencies by the solar X-flare.
The relevant ionosphere modification might be found in the following way. One picks a conductivity
profile and computes the set of SR spectra for realistic global distributions of thunderstorms during
the event. In these computations, a single model parameter is modified, and its selection depends on
some additional information or consideration. The set of SR spectra computed for the same source
distribution and the varying profile is compared with observations, so that a particular parameter
might be found by fitting the experimental data in the best way (e.g. [Nickolaenko and Hayakawa ,
2014]).
The exhausting search is relatively safe, but it is a time consuming approach. We briefly describe
here the rigorous formulation and solution of inverse SR problem aiming at the global distribution of
thunderstorms. We must remark that only an “effective” distribution might be found from the SR
data. Indeed, the basic resonance frequency is equal to 8 Hz while the pulse rate of the global
lightning is 50–100 events/s. One cannot resolve these pulses in a record and cannot formulate the
problem of electromagnetic location of ordinary lightning strokes.
The formal solution of the problem implies the Tikhonov regularization procedure to a linear
system of algebraic equations. One can present the links of the power spectra of the field components
in the form matrix relation:
Copy right @ESES
IJEAR
55
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
b = AL
(39)
Here b denotes the vector of measured power spectra, and L is the sought vector of the stroke rate at
different distances from the observer. The elements of A matrix contain the model spectra in the
particular Earth–ionosphere cavity that were computed for the set of distances from the vertical
electric dipole source as seen in Figure 22.
A
B
Figure 22: Amplitude spectra of vertical electric and horizontal magnetic field components for the SOD D
ranging from 1 to 19 Mm. To separate plots, we shifted them vertically by 0.25 D, so that the top curves
correspond to D = 1 Mm, and the bottom ones, to D = 19 Mm.
The SOD D ranges in Figure 25 from 1 Mm to 19 Mm varying with a 1 Mm step. Horizontal
axes show the frequency in Hz, and the vertical axes depict the spectral amplitude on logarithmic
scale. To separate different plots, we move them vertically by 0.25*D. The stroke current moment
was |МС()| = 108 A m (Hz)1/2 and the electric and magnetic spectral densities were measured in
mV/m and A/m correspondingly. The standard frequency dependence (f) was used specified by
Eqs. (13–16). One may observe that the regular succession of SR peaks is modulated by interference
of the direct and antipodal waves. This interference was responsible for the deep extended canyons
in the 2D maps of Figure 3. This feature is used to find the distance to the stroke that caused the Qburst: every distance has a unique spectral outline.
Copy right @ESES
IJEAR
56
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
The system (39) is resolved by the minimization of the squared deviations of the following
functional:
F(x, R)=||AL  b||2 + R ||L||2,
L  0,
(40)
where R is the regularization parameter [Shvets et al .,2010; Shvets et al ., 2011; Hayakawa et al .,
2014 ].
After finding vectors L (the stroke distribution along the distance from the observatory) by using
the data at a few observatories, one can compile a map of the global distribution of field sources
corresponding to the SR record. We illustrate the major stages of solution in Figure 23.
Figure 23: Major stages of solution of inverse problem based on simultaneous records at Moshiri, West
Greenwich, and Lehta that at UT=15 hr on January 1, 2000
Three global maps in the upper part of Figure 23 depict the lightning rate (events per second)
derived from the SR background spectra recorded simultaneously at Moshiri, West Greenwich, and
Lehta at 15 hr UT on Jan. 1, 2000. The rates are shown as the strips of equal distances from the
observatories. In the rectangular projection of the globe, the line of equal distance is a sinusoid, so
that distance distributions are seen as concentric curved color bands. The color scale relevant to the
maps is shown on the right. The distance profiles for every observatory allow constructing the
equivalent spatial distribution of lightning strokes over the globe. The simplest way to do this task is
to compute the geometric average, which is shown in the left lower frame of Figure 23. The map
indicates that global thunderstorms are concentrated at Africa, which is quite natural for the 15 hr UT
time moment. The higher activity is also apparent around the Ceylon Island and to the south of
Copy right @ESES
IJEAR
57
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Madagascar.
A more rigorous approach implies the tomographic reconstruction based on the set of spatial
distributions [Shvets et al ., 2010; Shvets et al ., 2011; Hayakawa et al ., 2014 ]. Results of
tomographic reconstruction are shown in the right lower panel of Fig. 23. Here, the crosses mark the
observatories, and the color inking corresponds to the lightning rate. As a whole, the geometric
averaging and the tomographic reconstruction provide similar geophysical results. However, the
rigorous procedure performs a substantial “cleaning” of the global thunderstorm distribution: it
removes insignificant and probably false objects while the most pronounced and compact areas are
preserved.
Apparently, the average annual distribution of lightning is most interesting, and we show it in
Figure 24 [Shvets et al., 2010; Shvets and Hayakawa, 2011; Hayakawa et al., 2011]. This map was
constructed similarly from the SR monitoring at the same three stations shown in the figure by
crosses. SR data cover the interval of August 1999 through May 2000. Here, the spatial distribution
of lightning rate was obtained by the tomographic reconstruction with an account of the azimuthal
data of the wave arrival (records of the orthogonal magnetic field components were used). The map
of the global distribution of the lightning flashes is shown for comparison in the lower part of Fig.
24, which was observed by the OTD spacecraft [Christian et al., 2003 ]. One may observe the
unmistakable similarity of optical orbital observations and the ground-based tomographic
reconstruction of global lightning activity. Actually, there are definite departures in the maps, which
indicate the necessity of further experiments and improvements.
The rigorously posed inverse problem of the SR is resolved in two stages. First, the power
spectra of magnetic and electric fields (Figure 22) are inversed giving the distance distributions of
thunderstorms relative to each observatory (Figure 23). Then, the distance profiles are applied as
tomographic projections in the recovery of the spatial distribution of sources (Figure 24). Accuracy
of solution increases when the wave arrival information is accounted for. Reconstruction based on
the records from Lehta (Russia), Moshiri (Japan), and West Greenwich (USA) showed the
diurnal/seasonal dynamics of the global distribution of lightning strokes, which fairly agree with the
data of meteorology and optical observations from space [Shvets and Hayakawa, 2011; Hayakawa et
al., 2011; Roldugin et al., 1999 ].
Copy right @ESES
IJEAR
58
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Figure 24: The yearly averaged (August 1999 – May 2000) global lightning distributions obtained from the
SR records and observed by OTD
The rigorously posed inverse problem of the SR is resolved in two stages. First, the power
spectra of magnetic and electric fields (Figure 22) are inversed giving the distance distributions of
thunderstorms relative to each observatory (Figure 23). Then, the distance profiles are applied as
tomographic projections in the recovery of the spatial distribution of sources (Figure 24). Accuracy
of solution increases when the wave arrival information is accounted for. Reconstruction based on
the records from Lehta (Russia), Moshiri (Japan), and West Greenwich (USA) showed the
diurnal/seasonal dynamics of the global distribution of lightning strokes, which fairly agree with the
data of meteorology and optical observations from space [Shvets and Hayakawa, 2011; Hayakawa et
al., 2011; Roldugin et al., 1999 ].
3.2.
Non-stationary ELF waves (Q-bursts)
The term of Q-bursts was introduced by Ogawa‟s group in 1966 [Ogawa et al., 1966], which refers
to discrete natural radio pulses that have the duration of 0.3 – 1 s and are detected all over the globe.
Copy right @ESES
IJEAR
59
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
The Q-bursts or ELF transients arrive from exclusively powerful lightning discharges and their
amplitude may exceed that of regular SR background by a factor of ten. These signals present a rare
possibility of study on the wave arriving from individual source, i.e., to derive the source position
and the radiation spectrum. The technique of locating the causative source was suggested in [Jones,
1970b; Burke and Jones, 1995; Burke and Jones, 1996 78–80], and it was developed and applied
in different studies, especially in investigations of transient luminous events (TLEs) including red
sprites.
We already demonstrated equations, the model spectra and the model pulsed waveforms
computed for different SODs. Here, we compare the model waveform computed directly in the time
domain by using Eq. (11) with the most advanced observations recorded in the extraordinary wide
frequency band. The experimental data were published in [Ogawa and Komatsu, 2007]. These
records practically do not contain the 50 Hz power supply interference, so that notch filtering was
unnecessary. The waveforms were recorded by the receiver of unusual bandwidth of 1 Hz–11 kHz,
and the sampling frequency was 22 kHz. The vertical electric field component was recorded at
Kochi, Japan (33.3 N, 133.4 E) during the years of 2003–2005. Individual pulses were selected
visually from the records performed in the fine weather conditions with the help of classical ball
antenna.
A comparison of model and observational data is made in Figure 25. The source distance was
evaluated as D = 16.5 Mm (see [Nickolaenko et al., 2008] for details). The waveform is shown by a
red curve with dots computed by using Eq. (11) for the linear frequency dependence of propagation
constant. The black line depicts the experimental waveform. One can note the remnants of the power
supply 50 Hz interference in the record. The fine structure is apparent in the record formed by the
spikes of individual lightning strokes, some of these having high amplitude. All these interesting and
convincing features are much smaller in comparison with the major excursion of the Q-burst.
The similarity of experimental and model is apparent in Figure 25. One can conclude that the
SOD was evaluated correctly, because the direct and the antipodal waves are well resolved and
coincident in the initial part of the record. The secondary round-the-world pulses are also clear and
coincident. The initial outline gave the name for the transient: the Q-burst of W–type. Owing to the
wave dispersion and attenuation of high frequencies, two pulses merge in the round-the-world wave.
It is remarkable that positions are coincident of the model and observed sub-pulses in the “head”, and
the same is valid for the observed and model round-the-world pulses.
Copy right @ESES
IJEAR
60
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Similarity of the model and observational waveforms validates the application of:
(i)
The uniform model of the Earth–ionosphere cavity bounded by an isotropic ionosphere;
(ii)
The propagation constant linearly varying with frequency;
(iii)
The formal time domain solution for the ELF transient.
Figure 25: Observed (black) and the model (red) waveforms of the Q-burst
Usually the SOD is evaluated from the spectral patterns of natural ELF pulses. It was shown in
[Jones, 1970b ] that the amplitude spectra of E- and H- field components have the unique outline
corresponding to the SOD as seen in Figure 22 where amplitude spectra of electric and magnetic
fields are shown.
Parametric ELF burst caused by the gamma flare on December 27, 2004
We already mentioned the impact of cosmic gamma flares on the background spectrum of the SR.
The huge abrupt modification of the dayside atmosphere by gamma rays has altered the leakage
current in the global Earth–ionosphere electric circuit. As a result, a parametric ELF radio pulse was
radiated, which was successfully detected by the SR observatories where we continuously monitor
the natural signal with the GPS time stamps [Nickolaenko and Hayakawa 2014 ; Tanaka et al., 2011 ;
Schekotov et al., 2013 ; Nickolaenko et al., 2014 ].
We present in Figure 25 the simultaneous experimental records of this pulse at the set
observatories. The field sites were: Moshiri, Japan (44.37N; 142.24E); Onagawa, Japan (38.43N;
Copy right @ESES
IJEAR
61
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
141.48E); Esrange, Sweden (67.83N; 21.1E); Karimshino, Russia (52.83N; 158.13E);
Nagycenk, Hungary (47.6N; 16.7E); and Hornsund, Spitzbergen (77.0N; 15.5E).
At the observatories of Esrange, Onagawa and Karimshino they recorded two orthogonal
horizontal magnetic field components, and at the sites at Moshiri, Nagycenk, and Hornsund they
registered two magnetic components and the vertical electric field. All records were supplied with
the GPS time stamps, so that time axis of individual records might deviate by a few milliseconds
utmost. Magnetic antennas at all observatories were oriented parallel to the geographic cardinal
directions, except the Karimshino observatory where they are positioned along the geomagnetic
coordinates. Owing to a rather small (6.5) magnetic declination at this site, the component D at
Karimshino practically coincides with the field HWE or HX component and the projection H is the HSN
or HY field. The vertical electric field is exceptionally important for establishing the source polarity
and finding the unambiguous source azimuth. Luckily, at many field sites we acquired the concurrent
ELF records in many channels, but none of them presented the reliable simultaneous records in all
three field components.
We show the collection of records made at six above-mentioned observatories in Figure 26. The
time (UT) is shown along the abscissa in the 1.5 s vicinity of 21:30:26.5 UT, which is the arrival
time of galactic gamma rays. The ordinates show the field amplitudes plotted in arbitrary units.
The upper plot depicts temporal variations of the vertical electric field at the Nagycenk
observatory. Two distinct pulses are clearly seen here, and since their onsets are negative, we can
state that both of them arrive from the source of positive polarity (the source current was directed
downward). Horizontal magnetic field records are assembled in pairs of all other sites. Before
plotting the data in Fig. 26, we had to remove the industrial noise with the help of the low-pass filters
and the Singular Spectral Analysis (SSA), or the Caterpillar procedure. Relevant details might be
found in [Nickolaenko and Hayakawa 2014 ]. This procedure removed both the 50 Hz interference
and the high frequency noise.
Figure 26 demonstrates outstanding similarity of data collected worldwide, although some
departures are obvious. Deviations originate from different propagation conditions and from the local
noise. Also, there are differences arising from deviations in particular frequency characteristics of
receiving equipment, however, we show the records as they were. Three thick black horizontal bars
at the time axis denote sections containing the distinctive pulse fragments: 26.3–26.4; 26.5–26.6; and
26.7–26.95 s. After applying the wavelet analysis to the waveforms of horizontal magnetic fields, we
Copy right @ESES
IJEAR
62
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
found relevant wave arrival angles (source azimuth AZ) and conducted the source triangulation from
the field sites for all these sections of the records. The azimuth AZ is the angle between the local
Figure 26: ELF records at different SR observatories. The (UT) is shown on the abscissa in the vicinity of the
onset of gamma flare.
south–north direction (Y-axis) and the major axis of the wave polarization ellipse. The azimuth is
counted clockwise, as a compass. Polarization characteristics were also computed through the
elements PYY = <HY HY*>, PXX = <HX HX*> and PXY = <HX HY*> of the coherence matrix, which were
derived from the spectra of the field components HX and HY.
Preliminary model computations indicated the expected features of parametric pulse and thus
facilitated the investigation. These were as follows:
Copy right @ESES
IJEAR
63
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
1. The parametric pulse is formed around the arrival time of gamma rays.
2. Owing to the great size of parametric source, radio signal contains only the lowest SR modes.
3. The source polarity must be positive, i.e., the pulse onset must be negative in the vertical
electric field.
4. Polarization of the pulsed magnetic field is linear, provided that modification of the global
circuit was symmetric.
5. The wave arrival azimuth must point to the center of modification (20.4S and 146.2W).
6. Amplitude of the pulse must noticeably exceed the SR background.
All these features are present in the records shown in Figure 26. The problem is that these are
pertinent to the couple of pulses: one occurred prior to the detection of gamma ray arrival, and the
second one took place at the right time. Analysis indicated that the first, precursory pulse satisfied all
demands except the early arrival time. The second pulse fitted all necessary characteristics, but the
source triangulation was highly unstable (scattered). It was concluded having in mind this feature
that the most probably both of these pulses were associated with the gamma flare.
A few words might be added concerning parametric excitation of SR transient signals in
connection with recent successful detection of gravitational waves at 09:50:45 UT on September 14,
2015 [Abbot, 2016 ]. Possible modulation of the Earth–ionosphere cavity by a gravitational wave
was modeled in [Minakov, et al., 1992 ]. It was shown that modification of SR by gravitational wave
will be most pronounced at the second or the third resonance modes, which is conditioned by the
geometry of modification and the finite propagation velocity. We will not go deeper in this issue and
mention only the general outline of the phenomenon. Let the plane gravitational wave be incident on
the Earth. The changes in the space-time metric will occur along the direction of propagation, so that
modification of the Earth–ionosphere cavity will be similar to the polar non-uniformity. It will match
the spatial field distribution of the second SR mode, so that the effect will reach its maximum at the
frequency of gravitational wave ~14 Hz. If one takes into account the delay due to the propagation
time of gravitational waves through the diameter of the Earth, it appears that the effect is enhanced at
the 20 Hz or at the third mode. A parametric pulse (if any) originating in the Earth–ionosphere cavity
from the gravitational modulation should occupy the band of 14 – 20 Hz. The gravitational wave
[Abbot, 2016] was detected above the 60 Hz with a sharp maximum around 130 Hz, which excludes
its possible detection by using SR. Thus, the Earth–ionosphere cavity was too large for an optimal
detection of transient gravitational wave of year 2015. The smaller planetary cavities might be
suitable for this task.
Copy right @ESES
IJEAR
64
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Distortion of signal waveform by a typical receiver
Computations of the pulse waveform at the output of a typical SR receiver showed that the signal
shape is significantly disturbed: the pulse onset is delayed, substantial changes are possible in the
amplitudes of positive and negative half-waves in the onset of pulsed signal, and the round-the-world
and antipodal pulses become poorly distinguishable. We demonstrate below that the initial pulsed
waveform might be partially restored when the phase modifications are compensated, which were
caused by the receiver [Yatsevich et al., 2014].
A typical SR receiver has the finite bandwidth (usually, 4–40 Hz) and a rejection filter of the
power supply frequency. Therefore, the pulsed waveforms might be modified beyond the
recognition. Substantial delay might appear reaching tens of milliseconds, which might lead to
wrong attribution when compared with other events. Changes in the positive and the negative
excursions in the pulse onset might cause the problems in establishing the source polarity, especially
in the presence of natural background resonance signal. An error in the arrival time of initial pulse
and the smoothed antipodal and round-the-world waves might increase an error in determining the
SOD. These features are important when analyzing the Q-burst records performed concurrently with
the optical observations of TLEs in the mesosphere above the powerful lightning strokes, the socalled "red sprites".
It is clear that some correction is desirable of the experimental pulsed waveforms at the output of
a typical SR receiver.
Changes in the waveform are mainly related to the phase characteristic of the receiver, and these
might be compensated during the consecutive processing or during direct recording of the pulse in
PC. The easiest way of compensating these distortions is the multiplication of the complex spectrum
of the signal at the output of the receiver by the complex conjugate of the receiver gain. Model
computations show that the inverse Fourier transform in this case will almost restore the correct
pulsed waveform (apart from modifications due to finite bandwidth). Such operation is equivalent to
the analog procedure of the “reverse playback”, in which one takes the signal recorded at the receiver
output, reverses it in the time, and sends it to the input of the same receiver. This procedure might be
done by the “backward” playback of the tape recorder. In both of these cases, the phase shifts
introduced by the receiver are completely compensated. Though the distortions remain that were
caused by the finite bandwidth, and these are observed as smoothing and widening of the time
domain signal. Moreover, the modification due to the receiver finite bandwidth will be enhanced
owing to the squaring of the amplitude gain of the receiver in such operations.
Copy right @ESES
IJEAR
65
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
The compensation procedure eliminates delays within the pulse and restores the original
amplitudes of the positive and negative half-waves in the pulse onset. As a result, more accurate and
fair assessment of the source parameters is obtained: the arrival time, the polarity, and the SOD. An
example of such a processing is shown in Figure 27. Here, we show the actual record of the vertical
electric field component of a Q–burst performed on January 21, 2001 at 22:54:12 UT at the Lehta
observatory. The sampling rate was 204 Hz. Time in milliseconds is plotted on the abscissa, and the
strength of vertical electric field is shown along the ordinates in arbitrary units. The red line in the
upper panel shows the EZ(t) signal recorded at the output of SR receiver. The blue line in the lower
frame depicts the same pulse after compensation of the receiver phase characteristic. We computed
the complex Fourier spectrum of the recorded pulse, multiplied it by the complex conjugate of the
known receiver gain, and performed afterwards the inverse Fourier transform, which returned the
records into the time domain. The through calibration of the complex gain of the receiver electric
channel was regularly made at the observatory with the help of a special radiating rod antenna.
Figure 27: Waveforms of a Q-burst with and without compensation of the receiver phase characteristic
As might be seen from Figure 27, the phase characteristic of the real receiver has caused a
significant (~40 ms) delay of the recorded pulse onset (red curve in upper plot) relative to its actual
initiation (blue curve in the lower plot). It is important that after compensation of the phase
Copy right @ESES
IJEAR
66
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
distortions one perceives that the observed Q-burst signal starts with the negative excursion. This
suggests that it is originated from the positive cloud-to-ground stroke. It is remarkable that the red
curve in Figure 27 started with the positive half-wave. Thus, the particular signal shape might be
altered by the equipment, which definitely would impede its application in geophysical studies.
Compensation of the receiver gain is strongly desirable.
4.
Detection of artificial ELF signals
The ELF radio propagation in the Earth-ionosphere cavity can also be studied by using ELF
transmissions. There are only two ELF transmitters in the world. The first one is the United States
WTF/MTF (Wisconsin Transmitter Facility/Michigan Transmitter Facility) having two crossed
transmitting antenna, and it operates at the center frequency of 46 and 76 Hz [ Bannister, 1999 ;
Fraser-Smith and Bannister, 1998]. The second one is the Russian Kola Peninsula Transmitter
(KPT, working at 82 Hz [Bannister, 1999;
Fraser-Smith and Bannister, 1998; Nickolaenko, 2008 ].
We present data on the Russian KPT transmissions detected in Japan, Antarctica and in the Atlantic
Ocean.
The center of KPT antenna is approximately located at the geographic coordinates (68.8ºN and
34.5ºE). As described in [Velikhov et al., 1996], the transmitter consists of two swept-frequency
generators of sinusoidal voltage and two parallel horizontal grounded wire antennas, each about 60
km long. The generator provides currents from 200 to 300 A in the antennas in the frequency range
from 20 to 250 Hz.
Observation on the Russian KPT transmissions detected in:
(i) Japan
Our data present the particular frequency of 82 Hz, and the ELF equipment was placed at
Moshiri in Japan. The detailed information might be found in [Yano et al., 2010]. We only note here
that the long-term data covering a whole year of 2007 indicated that the measured signal amplitude is
consistent with earlier results [Fraser-Smith. and Bannister, 1998]; the wave polarization was linear,
and the source bearing found by the goniometer technique agrees with the geometry of experiment.
Two horizontal magnetic fields HSN and HWE were detected by induction coil magnetometers and
the vertical electric field EZ was received by the „ball‟ antenna. The whole system was calibrated
before installation, so that the absolute amplitudes were obtained of the ELF fields. The sampling
Copy right @ESES
IJEAR
67
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
frequency was 4 kHz and the ELF waveforms were continuously recorded on the hard disk of the
PC. The relative position of observatory and of KPT corresponds to the wave arrival angle of 24.8º
from North to West, which was in fact the source bearing found in the measurements. The recoded
signals were processed by using FFT procedure with the segments of data 65.536 s long, so that
relevant frequency resolution was 0.01526 Hz. The amplitude spectra of each field component were
computed every 60 seconds, and thus we obtained dynamic spectra with temporal resolution of 1
minute [Yano et al., 2010].
After deriving the amplitude spectra |HSN(f)| and |HWE(f)|, the amplitude spectrum was found of
the total horizontal magnetic field H = [HSN2 + HWE2]1/2. Concurrently, the phase difference was
deduced of the two horizontal magnetic fields (f) = θSN(f) θWE(f) where θSN(f) = arg[HSN(f)] and
θWE(f) = arg[HWE(f)] are the arguments of the complex spectra HSN(f) and HWE(f) correspondingly.
The phase difference  determines the ELF wave polarization.
Figure 28 depicts the signal characteristics detected on the particular day of January 08, 2007.
The lowest panel here shows diurnal variations of the signal amplitude that reach 0.2 A/m. Intervals
of the man-made signal reception are marked by the boxes in the figure. One may note from the
amplitude data that the maximal signal-to-noise ratio was about 5.
Dots in the top panel of Figure 28 present the orientation of the complete magnetic field vector
φH. The angle φH  65 was found from the orthogonal components HSN and HWE, and it is connected
with the wave arrival angle φk by the relation φk=φH90º, and the wave arrival angle was found from
the formula: φk = tan1 [HSN(f)/HWE(f)]|f=
82 Hz.
Thus, the direction to the source was φk  25
(rotation from North to West), which is very close to the geometrical estimate. One may observe that
the arrival angle usually fluctuates due to the presence of natural ELF noise originating from the
global thunderstorms. When the man-made signal appears, the wave arrival azimuth becomes stable,
as seen in the boxes in the figure.
Points in the middle panel of Figure 28 depict the dynamic spectrum of the phase difference  =
(θWE  θSN). When the artificial signal is absent and only atmospheric noise is detected, the phase
difference covers the interval from –180 to +180 degrees. When the KPT signal appears, all points of
the plot concentrate around zero, thus indicating on the linear polarization of the arriving radio wave.
Thus, one can use the goniometer technique for establishing the source bearing.
Copy right @ESES
IJEAR
68
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Figure 28: Records of man-made ELF radio signal on 8 January, 2007
Detected amplitude of transmitted radio signal had the average value of 2.0107 A/m or 134
dB. The complex propagation constant (f) was estimated by using the characteristic heights hE and
hM of the two-scale exponential profile [Bannister, 1999 ] for the ambient day and night conditions:
hED= 55.1 km and hMD= 75.6 km; hEN= 75.6 km and hMN= 91.5 km. The propagation constant was
found from the standard equations pertinent to exponential profile. The wave attenuations rate was
D = 1.33 dB/Mm for the day and N = 0.82 dB/Mm for the night conditions. The phase velocity
(V) was characterized by (c/V)D = 1.245 and (c/V)N = 1.12 values. The model computations for the
experiment geometry and the cosine angular pattern of transmitting antenna has predicted the signal
amplitude of 137 dB at Moshiri (the distance of 6.2 Mm) [Yano et al., 2010], and its deviation from
the measured 134 dB was attributed to the underestimated current moment of the source
ML=1.421011 A/m2. This estimate is higher than that found in [Fraser-Smith. and Bannister, 1998]:
ML200(A)55(km)10(km) = 1.1105 (Akm)2.
An additional point is the particular SOD of 6 Mm, which corresponds to the critical distance
where the distance dependence intersects of the field amplitude in the ambient day and night
Copy right @ESES
IJEAR
69
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
conditions [Nickolaenko, 2008]. This is why no significant day–night variations were observed. The
physical explanation of the critical distance is rather simple. The wave attenuation is higher in the
day propagation, and relevant amplitude decreases with distance faster than in the ambient night
conditions. Simultaneously, the excitation factor of a source is greater in ambient day conditions.
Thus, the rapidly decreasing „day‟ curve starts from the higher initial amplitude, so it meets the
nighttime curve at the critical distance of about 6 Mm. The absence of regular daily variations might
play a positive role when the man-made signal is used to detect the seismogenic ionosphere
disturbances.
(ii) Antarctica
The so-called KPT is one of the sources of artificial ELF radiation whose signals were clearly
detected throughout the world [Fraser-Smith and Bannister, 1998]. A high amplitude signal was
noted reaching –142 dB at the 16.5 Mm distance detected at Dunedin, New Zealand. The level was –
151.9 dB at 18.3 Mm distance recorded at Arrival Heights, Antarctica.
We present the amplitude records performed at the “Bellinshausen” Soviet Antarctic station
(62.2S and 59W) in February 1989. It is well known that the radiating horizontal grounded wire
antenna acts as a vertical loop with the return current flowing through the Earth‟s crust. Such an
antenna is a horizontal magnetic dipole source. The maximum of its radiation pattern is found along
the direction of horizontal wire ground (the east–west direction in our case). With the satisfactory
accuracy one may apply the model of a vertical source current loop collinear to the geographic
parallel centered at 68.8N and 34.5E.
The “Bellinshausen” station is situated at the Waterloo (King George) Island in the Drake Strait
(62.2S and 59W). The SOD is equal to 16.25 Mm. Measurements were performed with one of two
antennas: vertical electric or the horizontal magnetic antenna, and we present the electric field data
here. The vertical electric antenna was a typical antenna with the 50 pF self-capacitance of active
electrode mounted on the top the metallic mast 3 m high. The input resistance of electric antenna
pre-amplifier was 109 Ohm.
The receiver of man-made signal was a narrow band RC filter designed as a combination of
active low-pass and the high-pass filters. It provided the bandwidth of 0.8 – 1.0 Hz centered at the
working frequency. The signal from this pre-selector was fed to the synchronous detector that
performed additional filtering. The two stage synchronous detector had the integration time of 50 s,
and it was locked-in with the high precision signal generator of 82.01 Hz frequency. Such a shift
causes 0.01 Hz amplitude modulation (sinusoidal beating) at the detector output, provided that the
Copy right @ESES
IJEAR
70
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
man-made radio signal arrives at the antenna. The amplitude of beating is directly proportional to
that of the input radio signal. The slow fluctuations are observed in the absence of artificial radiation.
The sensitivity of receiver was regularly checked and the effective height of the vertical electric
antenna was about 1 m in the Antarctic measurements [Nickolaenko, 2008].
The output of synchronous detector was sent to the multi–channel pen recorder, and the „beating‟
was clearly observed in the presence of man-made signal. Later, the paper chart was processed
manually, and the beating amplitude was measured, averaged over 10-min intervals and fixed in the
logbook.
Additional filtering and signal accumulation was performed simultaneously by the analog-digital
spectrum analyzer. For this purpose, the output of synchronous detector was sampled in the spectrum
analyzer with the sampling frequency of 0.25 Hz, and the instantaneous spectrum was obtained of
the relevant beating wave. The frequency channel of the man-made transmissions was also logged by
the chart record. Application of a spectrum analyzer allowed for extending the coherent
accumulation time to 40 minutes, so that the signal-to-noise ratio was increased up to 10 – 12.
Experimental data are presented in Figure 29. The UT is plotted on the abscissa here, and the
averaged amplitude variations are shown along the ordinate in the form of a histogram. The plot
presents the averaged data recorded from Feb. 07 to Feb 14, 1999. The amplitude of vertical electric
field is measured in dB relative to 5 V/m. One may observe that signal amplitude increases by
about 4 dB during the sunrise. It starts just before the local sunset, and the effect was noted in both
electric and magnetic fields. An increase in the electric and magnetic fields indicates that the effect
was conditioned by the decrease of wave attenuation on the propagation path. The relatively high
amplitude was recorded all through the local night, and it noticeably reduced just before the local
sunrise. Around the sunrise, the electric field amplitude showed a short, but distinct maximum, while
the field remained small during the local day gradually increasing toward the night. The magnetic
field amplitude behaves in a slightly different way: the narrow local peak in the amplitude was
absent after the sunrise.
We also show in Figure 29 the results of simplistic modelling. The line with dots was obtained in
the following way. According to the model [Nickolaenko, 2008 ] alteration in the wave attenuation
factor from day to night was equal to 0.51 dB/Mm. For the path length of 16.25 Mm, the alteration
increases the night amplitude by 8.3 dB in comparison with the day level. On the other hand, the
daytime ionosphere is lower than the night ionosphere, and changes in the field excitation factor will
work in the opposite direction: they will raise the daytime field by 2.3 dB. As a result, a transition
Copy right @ESES
IJEAR
71
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
from the ambient day to the ambient night propagation conditions will be associated with the 6 dB
filed increase. By choosing the daytime amplitude of 1 dB, we obtain the night level of 7 dB. These
two levels are shown by horizontal lines with dots in Fig. 29 linearly interpolated during the daynight transition period. One may observe that modelling agrees with observations.
The solid blue line in Figure 29 depicts the results of alternative interpretation. Now, the
deviations in the field excitation factor are ignored, but the temporal changes at the day and night
segments on the propagation path are more accurately accounted. The coordinates of the point T(k,
k) are found on the propagation path where the sunrise (sunset) is observed at the given moment of
universal time. This point divides the path into the “day” and the “night” portions, and the relevant
field amplitude yk in dB is calculate by using the following formula:
yk = 1.33dD0.82dN
(41)
where dD and dN are the day and night portions of propagation path. The relevant amplitude
variation is shown by the solid blue line in Figure 29 deviating from the linear interpolation. The
range of amplitude variations increases owing to a disregard of modulations in the excitation factors.
Figure 29 : Diurnal variations of man-made signal in Antarctica
The above explanations of experimental data account for the position of the day – night interface
(solar terminator) that plays the key role. The localized peak around the sunrise time is definitely
linked to the “gliding propagation” of the ELF radio wave when the day–night interface and the arc
connecting the source and observer are almost coincident. Maps in Figure 30 indicate that the source
Copy right @ESES
IJEAR
72
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
and observer are positioned nearby the solar terminator line when the field increases at the
observatory. Such behavior suggests that the day–night interface is rather sharp. The transmitted
radio wave is incident on the day – night interface with a gliding angle at UT= 7 – 8 hr, so that one
can expect relatively high reflection at this time and a summation of the direct and reflected waves.
The absence of the peak in the magnetic field component night be explained by the orientation of
direct and mirror interfering vectors.
Figure 30: Global map with propagation path and the first Fresnel zone. Positions of morning (left frame) and
evening ((right frame) terminators are shown for different moments of UT. Time is printed on terminator
lines.
The habitual terms “propagation path” and “wave reflection from the day–night interface” must
be used at ELF with caution. One has to draw the “the radio beam” first, i.e. the first Fresnel zone as
seen in Figure 30. Distant source and the receiver are positioned in the Northern and Southern
hemispheres, and the first Fresnel zone is superimposed on the global map in the rectangular
coordinates. It replaces the customary propagation path at ELF, and we show such a path (the first
Fresnel zone) for the signal frequency of 4.1 kHz by the narrow yellow contour embracing the
geodetic source – observer arc. This high frequency radio beam still has a considerable width of ~4,
so that the terminator line crossed it in 15 minutes. Owing to a great path length and to the extremely
low frequency, the first Fresnel zone covers the 110 longitude interval at 82 Hz, so that Africa and
South America are found within this zone.
The morning terminator lines are shown on the ground surface in the left frame of Figure 30 for a
set of sunrise time (UT) printed for February 10, 1989. Obviously, these lines are practically
“parallel” to the source–observer arc. The right frame depicts geodetic lines of the evening
terminator. When UT = 7.5 hr, the sun has already risen at the transmitter, and the point of the
sunrise is found close to the Greenwich meridian at the propagation path. The morning terminator
Copy right @ESES
IJEAR
73
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
line moves across the source–observer arc rather fast, about one hour. In contrast, the evening
terminator (the right frame of figure) is almost perpendicular to this arc, and it crosses the source–
observer arc during 11 hr. Thus, the February sunrise occurs almost simultaneously at the points of
source–observer geodetic line, in spite the longitudinal separation of transmitter and observer
exceeds 90.
Sunrise and sunset are shown in Figure 29 by blue wedge and red triangle. The sunrise occurred
at the transmitter around 6.7 hr UT, and the sunset takes place on 13.5 hr. Despite the wide Fresnel
zone, the amplitude records of artificial ELF radio wave demonstrate the effect of terminator in an
interval of UT from 6 to 9 hr. It looks like that the effect emerged from the great SOD combined
with practically parallel positioning of the day–night interface and the source–observer geodetic line.
Approximately the observed amplitude increase reached 100% (6–8 dB) at the 82 Hz signal
frequency. If one assumes that the magnitude of terminator effect linearly varies with frequency, it
must decrease to 10% at the first SR mode. This estimate agrees rather well with the rigorous model
computations. Thus, observations of the man-made ELF radio signal correspond to those collected in
the SR observations.
(iii) Atlantic Ocean
The vertical ELF field component from the KPT was recorded on board the “Professor Zubov”
research vessel in the Atlantic Ocean from April to May 1989 during the return of 34 th Soviet
Antarctic expedition to the city of Leningrad [Nickolaenko, 2008 ].. The receiving vertical electric
antenna was installed at the port side of navigation deck of the ship. The same antenna received the
signals of the “Omega” VLF radio navigation system at 10.2 kHz, which allowed them to calibrate
its effective height and obtain heff = 1.1 0.1 m. Thus the absolute values of the ELF field could be
Figure 31: Field amplitude versus the SOD
obtained.
Copy right @ESES
IJEAR
74
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Measurements in the sea were made in the same way and with the same equipment as at the
“Bellinshausen” Antarctic station [Nickolaenko, 2008]. The recorded field amplitude was averaged
over the whole day, and this value was used as the function of the daily averaged SOD. The distance
was calculated for the vessel coordinates for the the ship‟s noon. Figure 31 depicts the distance
dependence of the observed field amplitude. The abscissa shows the distance from the vessel to the
center of radiating antenna in Mm. The ordinate depicts the field amplitude in dB relative to 1 V/m.
The dotted curve shows the experimental data, and the smooth red line depicts the computed
amplitude.
The ship coordinates during the measurements varied in the way that provided practical
invariability of the “launching angle” at the transmitting antenna: variations of azimuth of observer
seen at the transmitter were practically absent in the data set. Distance variations of the field
amplitude indicated that the wave attenuation rate was Im[ ]|f = 82 Hz = 0.9225 Napier/radian. The
estimate for the attenuation rate was obtained in the distance interval 1–8 Mm. The red line in
Figure 31 shows the model computation of the field amplitude for the above attenuation rate with an
account of slight alterations of the receiver azimuth relative the middle of radiating antenna.
Obviously the model data are in excellent agreement with observations.
5.
Tweek atmospherics (sferics)
Internal contradiction of experimental studies of transverse resonance: the resonance cannot be
observed at a short distance from the source – no propagating waves. Observations at some distance
suppose longitudinal propagation in the Earth-ionosphere duct at frequencies just above the modal
cutoff. When the SOD substantially exceeds the waveguide effective height, the pulsed signal
transforms into a tweek atmospheric (sferic). The following formula contains actually two
independent branches of eigen-values.
knp 
n n  1  p 


 b  a
a2
2
(42)
The first one is the SR condition pertinent to TM-waves that corresponds to p=0. The second
branch of the solution corresponds to p=1, 2, 3,… It describes the transverse resonance with
characteristic frequencies corresponding to the condition k (b–a)  p. The second term under the
square root in Eq. (42) becomes principal in this case, and we have the following expression for the
transverse resonance frequency,
Copy right @ESES
IJEAR
75
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
f np 
c p  1 n n  1
1 
2 b  a   2  p 2

2
 b  a  


 a  

ISSN 2395 0064
(43)
which practically gives the relation kpn(b–a) = p for an arbitrary n.
The basic transverse resonance (p=1) is found around 1.5-2.5 kHz frequency and is observed
predominantly at night [Rafalsky , 1991 ; Shvets, 1994; Hayakawa et al., 1994]. Indeed for the
ionospheric height of 100 km, we have f10 = 1.5 kHz with the corrections proportional to n(n+1) term
of approximately 1 Hz value [Hynninen, 1974]. This transverse resonance is a local phenomenon
because of relatively high attenuation factor of about 20 dB/Mm at its frequencies. Such waves do
not propagate over long distances from the sources. The frequencies (43) are regarded in the
literature as the cut-off frequencies of the Earth-ionosphere waveguide [Wait, 1970; Galejs, 1972;
Sukhorukov et al., 1992 ] because the p-th mode does not propagate at frequencies below fpn. Still,
special investigations had demonstrated that electromagnetic waves are indeed trapped in the
transverse cross-section of the Earth-ionosphere guiding system [Rafalsky , 1991 ; Shvets, 1994].
5.1.
Tweeks as a diagnostic tool for the lower ionosphere
Tweek atmospherics (tweeks) are electromagnetic pulses propagating in the natural Earth–ionosphere
waveguide, and they represent a response of the natural cavity to the excitation by lightning discharges.
Tweeks are formed due to small losses during night time in the ionosphere at altitudes from about 85 to
95 km, which effectively reflects ELF-VLF electromagnetic waves. A study of this altitude range in
the ionosphere meets essential difficulties due to relatively small electron densities (10–1,000 cm–3 ).
The sources of probing radiation are both the transmitters of VLF navigation and communication
systems, and the strokes of lightning. The amplitude and phase of narrow band signals from a VLF
transmitter vary with changes of the waveguide height, the slope of electron density profile, with the
irregularities appearing in the lower ionosphere close to the propagation path. Owing to the small VLF
attenuation rate in the ambient night conditions and at the distances smaller than 3,000 km, significant
contribution to the observed field concurrently with the basic normal wave (mode) is made by the
modes of higher orders. This detail hampers substantially the solution of the inverse problem when
deducing the waveguide parameters from the results of measurements of the narrow band radiation of
VLF radio stations, so that tweeks are considered to be a useful natural means for radio sounding the
lower ionosphere in a wide frequency band.
Copy right @ESES
IJEAR
76
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Tweeks were distinguished from the so-called “musical atmospheric disturbances” [Eckersley,
1925] as a separate phenomenon by Burton and Boardman [Burton and Boardman, 1933] who
extensively studied interferences on submarine cables. Having a duration less than 1/8 second these
distinct musical atmospherics were given their onomatopoeic name [Burton and Boardman, 1933].
“Tweeks practically always start above the 2000 Hz and reduce very rapidly toward a lower limiting
frequency”. An example of tweek records is presented in Figure 32.
Figure 32: Early records of tweeks (oscillogram of tweeks) and the timing impulse frequency, 1000 Hz.
Adapted from [Burton and Boardman, 1933].
Those authors explained the frequency change of tweeks based on the theory of multiple reflections
of the pulse emitted by a lightning stroke from the Earth and from the ionosphere. The approach (as
seen in Figure 33) was originally proposed by Barkhausen [Barkhausen, 1930] in an attempt to explain
the formation of whistling atmospherics: “whistlers”. Observations [Burton and Boardman, 1933]
showed that whistlers do not have a lower cutoff frequency, while the analysis of tweek atmospherics
revealed such a limit.
Figure 33: Multiple reflections of a lightning stroke pulse from the Earth and ionosphere. Adapted from
[Barkhausen, 1930].
The analysis of tape recordings of tweeks was made by Burton and Boardman [Burton and
Boardman, 1933] with the help of a sound spectrograph, and it revealed the presence of three distinct
harmonics [Potter, 1951]. Lynn and Crouchley [Lynn and Crouchley, 1967] have demonstrated that
the tweeks containing a few harmonics arrive from the sources at distances up to 3,000 km.
Copy right @ESES
IJEAR
77
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Owing to the reduced attenuation in the ionosphere, the spectrum of such signals increases nearby
the cutoff frequencies of the Earth–ionosphere waveguide, and this allows us to estimate the reflection
height and some other parameters of the lower ionosphere [Burton and Boardman, 1933; Reeve and
Rycroft , 1972; Mikhailova and Kapustina, 1988; Yedemsky et al., 1992 ; Ryabov, 1994; Shvets
and Hayakawa, 1998].
It was shown originally in [Singh and Singh, 1996 ] that the cutoff frequency of the Earth–
ionosphere waveguide evaluated directly from the higher harmonics of a tweek deviate from the
multiples of the cutoff frequency of the first (fundamental) harmonic. In particular, the proportionality
coefficients of the cutoff frequencies for the second and third modes were a few percent higher than the
corresponding integers. So the lower ionosphere edge cannot be described by a sharp boundary when
one treats the propagation of wideband ELF–VLF radio signals.
In turn, measurements of the cutoff frequencies of several modes in a tweek provide an opportunity
for evaluating parameters of the conductivity profile of the lower ionosphere [Shvets et al., 2014].
5.2.
Basic equations
To calculate field components from a vertical electric dipole it is necessary to sum up contributions
from all waveguide modes propagating at a given frequency. The following equations give the vertical
electric and horizontal magnetic field components on the Earth‟s surface from the dipole source with
the source current moment MC() at a distance D [Wait, 1970]:
Er ( )  i
H  ( )  i
H  ( )  i
M C ( ) exp  i  4


(44)


(45)


(46)
exp  ikS p D

12
pS p
exp  ikS p D
C  
12
pS p
exp  ikS p D
p 0
M C ( ) exp  i  4
h a sin D a
h a sin D a

32
pS p
h a sin D a
M C ( ) exp  i  4


p 0

p p
p 0
where  is the free space impedance. To calculate values of Sp = (1 – Cp2)1/2 for the perfectly
conducting Earth and anisotropic ionosphere it is necessary to obtain the solution of the following
dispersion equation [Wait, 1970; Budden, 1952 ],
W  0,



W  ei 2khC || R|| ei 2khC   R || R  R|| ,


(47)
dW 
where the excitation factor is determined by  p  i 2khC p e i 2khCp   R e i 2khCp 

 dC  C C p
, and the
reflection coefficient matrix R was calculated for homogeneous and inhomogeneous models of
Copy right @ESES
IJEAR
78
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
anisotropic ionosphere with quasi-longitudinal or arbitrary [3, 111] direction of the Earth‟s magnetic
field. The polarization factor of a particular mode p  || R
R


e
i 2khCp
 is the ratio of the complex
amplitudes of the horizontal components of the magnetic field in and perpendicular to the plane of
incidence [Budden, 1952 ].
Approximate solutions for tweek fields near cutoff frequencies were obtained in [Sukhorukov et al.,
1992; Ryabov, 1994; Sukhorukov, 1992; Sukhorukov, 1996] that allowed us to obtain analytical
expressions connecting the parameters of the lower ionosphere with peculiarities in the spectra of
tweeks.
5.3.
Methods for tweek analysis
The first analysis was aimed at the distance from the lightning stroke and the effective height of the
reflecting layer in the ionosphere, and the time–frequency variation was used of the individual tweek
[Burton and Boardman, 1933]. The instantaneous frequency of the signal was determined as the
reciprocal of the time delay between two successive peaks in the signal waveform fixed by an
oscillograph. The results of frequency determination (curves A and B) are presented in Figure 34,
where also the curve C calculated based on the multiple reflection theory [Barkhausen, 1930] is shown
for the distance of 1770 km and ionosphere height of 83.5 km.
To determine the parameters of the waveguide by using the tweek signal, it is necessary to know
the coordinates of the source, and these might be obtained from the independent networks of lightning
location [Cummer and Bell, 1998; Rodger et al., 2005]. However, this information is provided on a
commercial basis, and the realization of the lightning location network is technically difficult and
expensive. Therefore, the single-position techniques are used that allow for locating the parent
lightning strokes from the analysis of tweeks.
A method based on measurements of two orthogonal magnetic and vertical electric component of
the tweek field was proposed by Rafalsky et al. [Rafalsky et al ., 1995a ; Rafalsky et al ., 1995b ] and
termed later as “Kharkov method” [Rafalsky et al ., 1995b]. The phase spectrum of a single mode
related to the time of the signal reception is determined by kD(Sn – 1), where k is again the wave
number in free space, D is the SOD, and Sn is again the complex sine of the modal incidence angle.
While the waves propagating under grazing incidence angles are of little utility to estimate the
waveguide height, the frequency band near a cutoff frequency should be chosen. The single TE1 mode
can be isolated in the frequency band between the first and the second order mode cutoffs in the
Copy right @ESES
IJEAR
79
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
spectrum of the longitudinal magnetic component. The source azimuth is determined by finding
direction of the Poynting vector averaged over a wide frequency range and the longitudinal magnetic
component is extracted by the rotary coordinate system transformation. In the model of the perfectly
conducting plane waveguide the phase spectrum of the TE1 mode is determined as follows: F1 =
kD{[1-( /kh)2]1/2 – 1} and the inverse problem solution is reduced to fitting F1 to an experimental
phase spectrum by choosing the two parameters h and D. The method was validated in [Brundell et al
., 2002] by a comparison with results of lightning location using a network of Lightning Positioning
and Tracking System (LPATS) with sub kilometer location accuracy. The accuracy of the distance
finding of the “Kharkov method” was estimated to be 74 km for distances at least greater than 1,000
km.
Figure 34: A and B, tweek frequency variation curves, and C, computed curve. Adapted from [Burton
and Boardman, 1933].
Another traditional approach consists in measuring the time delay between the waves of different
frequencies pertinent to the same tweek, and the retard arises from the frequency dispersion in the
waveguide [Outsu, 1960; Ohya et al ., 2008 ]. By using the delay measured at several frequencies,
one can estimate both the distance from the source and the height of the waveguide.
Deducing the waveguide parameters for the higher-order modes presents a significant problem,
since the phase and amplitude spectra as a whole are formed by interference of several waveguide
modes. Resting on the Barkhausen‟s multiple reflection theory we can simply obtain an analytical
expression for the instant frequency in a tweek. It is obvious from Figure 33 that the propagation time
for the path with n reflections from the ionosphere is equal to,
Copy right @ESES
IJEAR
80
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
tn 
ISSN 2395 0064
1
D 2  2nh 2 , n  0,1,2,
c
(48)
where h = b – a is the effective height of the waveguide and D is the distance between the lightning
and observer. The instant frequency is defined by derivative tn by n from (48) [Yano et al ., 1989]:
c  D
f 

1   
dtn dn 2h   ct 
1
2
1 / 2


,
(49)
where t is the time counted from the moment of lightning origin.
This expression is obtained also from the frequency dispersion relation for the group velocity of
the first order waveguide mode [Hayakawa et al., 1995]. The waveguide mode theory yields that the
value fc = c/(2h) in Eq. (49) is the cutoff frequency of the first order mode. The strong frequency
dispersion of tweek harmonics is associated with a sharp reduction in the group velocity of radio waves
propagating in the Earth–ionosphere waveguide when the signal frequency approaches the cut-off
frequency of the waveguide. The cutoff frequency for higher order modes can be obtained by
dividing the waveguide height by the mode number p:
f cp 
pc
2h
(50)
Let  = t – D/c be the time from the moment of arrival of the direct ray to the observer, then the
instant frequency of Eq. (49) for the pth order mode transforms to the following expression:
  D 2 
f    f cp 1  
 
  D  c  
1 / 2
,
(51)
In spite of simplicity of the model this formula proves to be useful in tweek analysis to find both
the source range and effective waveguide height along the propagation path of a tweek by fitting the
theoretical dependence (51) to experimental one till present time [Shvets, et al, 2014; Ohya et al .,
2008 ]. In particular, when analyzing higher order harmonics in tweeks it becomes possible to estimate
the frequency dependence of the effective reflection altitude and in such a way, parameters of the
vertical conductivity profile of the lower ionosphere. An example of analyses of experimental tweek
records [Budden, 1952] is demonstrated in Figure . 35.
Copy right @ESES
IJEAR
81
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
a)
b)
ISSN 2395 0064
c)
d)
Figure 35 : Dynamic spectrum (a), waveform of the longitudinal magnetic component (b), its amplitude
spectrum (c) and estimated effective heights (d) for different waveguide modes for the tweek recorded at 5.55º
E, 16.7º S on January 21, 1991, at 20:29:22 UT. The estimated distance to the source and parameters of the
exponential ionospheric conductivity profile are as follows: D=1190 km, h1 =86.1 km, ζ0 =0.44 km. The circle
points with the standard deviation bars in (d) denote effective heights experimentally found, and the solid line is
the fitted theoretical dependence of h1 from Eqs. (22, 23) which yields the estimated parameters h1 and ζ0.
6.
Other ELF related topics
This review has dealt with the main parts of subionospheric ELF radio wave propagation
(fundamental theory, ELF natural noise (stationary and non-stationary), ELF transmitter signals, and
their use to deduce different geophysical information such as space weather effects etc), but many
other related topics are still existing, so we will list only a few below.
Other important directions are extensively pursued recently with the use of ELF waves treated in
this review: studies of TLEs in the mesosphere originating from the powerful positive lightning
discharges, because the whole view of TLEs is closely related to the most fundamental physics of
lightning (charge generation, acceleration of particles by the electric field etc.). This topic is also
addressed in the monographs [Nickolaenko and Hayakawa, 2014; Surkov, and Hayakawa, 2014].
Finally, we must also mention the attempts of detecting and applying the global electromagnetic
resonance on other planets of the solar system (Venus, Mars, Titan etc) and implications of the use
of SR data in detecting the environmental modifications by seismic and pre-seismic activity
dedicated to earthquake prediction [Hayakawa, 2015; Molchanov and Hayakawa 2008]. These
issues are exceptionally interesting, but they are beyond the scope of this brief review.
Copy right @ESES
IJEAR
82
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
References
Abbot, B.P. et al., Observations of gravitational waves from binary black hole merger, Phys. Rev.
Lett., 116, 061102, 16, 2016.
Ayurov, D., and Bashkuev, Yu. B., Schumann resonances in high latitudes,
Symposium, Kharkov, June 2016.
MSMW 2016
Bannister, P. R., Further examples of seasonal variations of ELF radio propagation parameters,
Radio Sci., 34, No.1, 199–208, 1999.
Barkhausen, H., Whistling tones from the Earth, Proc. IRE, 18, 1,155–1,159, 1930.
Bliokh, P.V., Nickolaenko, A.P., and Filippov, Yu. F., Schumann Resonances in the
Earthionosphere Cavity, Peter Perigrinus, London, 1980.
Brundell, J.B., Rodger, C.J., and Dowden, R.L., Validation of single station lightning location
technique, Radio Sci., 37. No 4, 1059-1067, doi: 10.1029/2001RS002477, 2002.
Budden, K. G. , The propagation of a radio atmospheric---II, Phil. Mag., 43, 1179, 1952.
Burke, C. P. and Jones, D. Ll., Global radiolocation in the lower ELF frequency band, J. Geophys.
Res., 100, 26,263 – 26, 271, 1995.
Burke, C.P. and Jones, D. Ll., On the polarity and continuing currents in unusually large lightning
flashes deduced from ELF events, J. Atmos. Terr. Phys., 58, 531540, 1996.
Burrows, M.L., ELF Communications Antennas, Peter Peregrinus, London, 245, 1978.
Burton, E.T. and Boardman, E.M., Audio-frequency atmospherics, Proc. IRE, 21, 1,476–1,494,
1933.
Burton, E.T. and Boardman, E.M., Effects of solar eclipse on audio frequency atmospherics, Nature,
131, 81-82, 1933.
Christian, H. J., Blakeslee, R. J., Boccippio, D. J., Boeck, W. L., Buechler, D. E., Driscoll, K. T.,
Goodman, S. J., Hall, J. M., Koshak, W. J., Mach, D. M. and Stewart, M. F., Global frequency and
distribution of lightning as observed from space by the Optical Transient Detector, J. Geophys. Res.,
108(D1), 4005, doi:10.1029/2002JD002347, 2003.
Cole, R.K. and Pierce, E.T., Electrification in the Earth‟s atmosphere from altitudes between 0 and
100 kilometers. J. Geophys. Res., 70 (11), 2,7352,749, 1965.
Cummer, S.A., Inan, U.S., and Bell, T.F., Ionospheric D-region remote sensing using VLF radio
atmospherics, Radio Sci., 33, No. 6, 1,781-1,792, 1998.
Eckersley, T. L., Note on musical atmospheric disturbances, Phil. Mag., 49 (5), 1,250–1,259, 1925.
Copy right @ESES
IJEAR
83
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Fraser-Smith, A.C. and Bannister, P. R. , Reception of ELF signals at antipodal distances, Radio
Sci., 33, 8388, 1998.
Füllekrug, M., Dispersion relation for spherical electromagnetic resonances in the atmosphere, Phys.
Lett. A, 275, 80 – 89, 2000.
Galejs, J., Terrestrial Propagation of Long Electromagnetic Waves, Pergamon Press, New York, 362,
1972.
Galuk, Yu.P., Nickolaenko, A.P., and Hayakawa, M., Comparison of exact and approximate
solutions of the Schumann resonance problem for the knee conductivity profile, Telecommun. Radio
Eng., 74 (15), 1,377-1,390, 2015.
Gradstein, C. and Ryzhik, I.M., Tables of Integrals, Sums, Series and Products, Phys-Math Publ.,
Moscow (in Russian), 1963.
Greifinger, C. and Greifinger, P., Approximate method for determining ELF eigenvalues in the
Earthionosphere waveguide, Radio Sci., 13, 831837, 1978.
Harth, W.,Theory of low frequency wave propagation, in Handbook of Atmospherics, Ed. by H.
Volland, Vol. 2, CRC Press, Boca Raton, FL, 133–202, 1982.
Hayakawa, M., Earthquake Prediction with Radio Techniques, John Wiley & Sons, Singapore, 289,
2015.
Hayakawa, M., Nickolaenko, A. P., Shvets, A. V. and Hobara, Y., Recent studies of Schumann
resonance and ELF transients, in Lightning: Properties, Formation and Types, Ed by M.D. Wood,
Nova Sci Pub., Ch. 3, pp.39 – 71, 2011.
Hayakawa, M., Ohta, K., and Baba, K., Wave characteristics of tweek atmospherics deduced from
the direction finding measurement and theoretical interpretation, J. Geophys. Res., 99, 10,73310,
743, 1994.
Hayakawa, M., Ohta, K., Shimakura, S. and Baba, K., Recent findings on VLF/ELF spherics, J.
Atmos. Terr. Phys., 57, 5, 467-477, 1995.
http://www.interscience.wiley.com:83/eeee/57/1257/W.1257-toc.html.
Huang, E., Williams, E., Boldi, R., Heckman, S., Lyons, W., Taylor, M., Nelson T., and Wong, C.,
Criteria for sprites and elves based on Schumann resonance observations, J. Geophys. Res., 104,
16,94316, 964, 1999.
Hutchins, M. L., Holzworth, R. H., Brundell, J. B. and Rodger, C. J., Relative detection efficiency of
the World Wide Lightning Location Network, Radio Sci., 47, RS6005, doi:10.1029/2012RS005049,
2012.
Copy right @ESES
IJEAR
84
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Hynninen, E.M., On the high-frequency resonance oscillations of electromagnetic field in the
Earthionosphere cavity, in Problems of Wave Diffraction and Propagation, 21, Leningrad State
University Press, 150154 (in Russian), 1974.
Hynninen, E.M. and Galuk, Yu.P. , Field of vertical dipole over the spherical Earth with nonuniform along height ionosphere, in Problems of Diffraction and Radio Wave Propagation, Vol. 11,
Leningrad State University Press, Leningrad, 90–120 (in Russian) ,1972.
Inan, U. S., Lehtinen, N. G., Moore, R. C., Hurley, K., Boggs, S., Smith, D. M. and Fishman, G. J.,
Massive disturbance of the daytime lower ionosphere by the giant  –ray flare from magnetar SGR
1806–20, Geophys. Res. Lett., 34, L08103, doi:10.1029/2006GL029145, 2007.
Ishaq, M. and Jones, D. Ll., Method of obtaining radiowave propagation parameters for the Earth–
ionosphere duct at ELF, Electronic Letters, 13(2), 254–255, 1977.
Jones, D. Ll., The apparent resonance frequencies of the Earthionosphere cavity when excited by a
single dipole source, J. Geomagn. Geoelectr., 21, 679684, 1969.
Jones, D. Ll., Numerical computations of terrestrial ELF electromagnetic wave fields in the
frequency domain, Radio Sci., 5, 803 – 809, 1970.
Jones, D. Ll., Propagation of ELF pulses in the Earth-ionosphere cavity and application to slow tail
sferics, Radio Sci., 5, 1,1531,163, 1970.
Jones, D. Ll. and Knott, M., Comparison of simplified and full-wave ELF propagation models,
Abstracts of Reports at XXVI URSI General Assembly, Section E6-10, Toronto, 295, 1999.
Kudintseva, I.G., Nickolaenko, A.P., Rycroft, M.J., and Odzimek, A., AC and DC global electric
circuit properties and the height profile of atmospheric conductivity, Ann. Geophys., in press, 2016.
Lynn, K.J.W., and Crouchley, J., Night-time spheric propagation at frequencies below 10 kHz, Aust.
J. Phys., 20, 101–108, 1967.
Madden, T. and Thompson, W., Low frequency electromagnetic oscillations of the earthionosphere
cavity, Rev. Geophys., 3, 211254, 1965.
Melnikov, A., Price, C., Sátori, G., and Füllekrug, M., Influence of solar terminator passages on
Schumann resonance parameters, J. Atmos. Solar-Terr. Phys., 66, doi:10.1016/j.jastp.2004.05.014,
2004.
Mikhailova, G.A., and Kapustina, O.V., The frequency-time structure of “tweek”-type atmospherics
and VLF diagnostics of parameters of the night-time lower ionosphere, Geomagn. Aeron., 28, 879–
882, 1988.
Copy right @ESES
IJEAR
85
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Minakov, A.A., Nickolaenko, A. P., and Rabinowicz, L. M. , Transformation of gravitational waves
into electromagnetic by the electrostatic field in the Earth-ionosphere cavity, Izvestija vuzov,
Radiofizika, 35, No. 6-7, 488-497, 1992, (in Russian). [Translated into English: Minakov, A.A.,
Nickolaenko, A.P., Rabinowicz, L.M., Gravitational-to-electromagnetic wave conversion in
electrostatic field of earth-ionosphere resonator, Radiophysics and Quantum Electronics, Vol. 35,
No. 6, 318-323, 1992.
Molchanov, O.A., and Hayakawa, M., Seismo Electromagnetics and Related Phenomena: Histroty
and latest results, TERRAPUB, Tokyo, 189, 2008.
Mushtak, V.C., and Williams, E., Propagation parameters for uniform models of the Earthionosphere waveguide, J. Atmos. Solar-Terr. Phys., 64(6), 1,989–2,001, 2002.
Nelson, P. H., Ionospheric perturbations and Schumann resonance data, Ph.D. Thesis, MIT,
Cambridge Mass, 1967.
Nickolaenko, A. and Hayakawa, M., Schumann Resonance for Tyros (Essentials of Global
Electromagnetic Resonance in the Earth–Ionosphere Cavity), Springer, Springer Geophysics Series
XI, Tokyo, 348, 2014.
Nickolaenko, A. P., and Hayakawa, M., Resonances in the Earth-ionosphere Cavity, Kluwer
Academic Publ., Dordrecht, 380, 2002.
Nickolaenko, A. P., Koloskov, A.V., Hayakawa, M., Yampolski, Yu.M., Budanov, O.V., and
Korepanov, V.E. 11-year solar cycle in Schumann resonance data as observed in Antarctica, Sun and
Geosphere, 10, No. 1, 39–49, 2015.
Nickolaenko, A. P., Pechony, O., and Price, C., Model variations of Schumann resonance based on
Optical Transient Detector maps of global lightning activity, J. Geophys. Res., 111, D23102,
doi:10.1029/2005JD006844, 2006b.
Nickolaenko, A. P., Schekotov, A. Yu., Hayakawa, M., Hobara, Y., Satori, G., Bor, J., and Neska,
M., Multi-point detection of the ELF transient caused by the gamma flare of December 27, 2004,
Radiophysics and Quantum Electronics, 57, No. 2, 125–140, 2014.
Nickolaenko, A. P., Yatsevich, E. I., Shvets, A. V., Hayakawa, M. and Hobara, Y., Universal and
local time variations deduced from simultaneous Schumann resonance records at three widely
separated observatories, Radio Sci., 46, RS5003, doi:10.1029/2011RS004663, 2011.
Nickolaenko, A.P., ELF attenuation factor derived from distance dependence of radio wave
amplitude propagating from an artificial source, Telecommun. Radio Eng., 67, No. 18, 1621 – 1629,
2008.
Nickolaenko, A.P., Impact of the Ionosphere day-night non-uniformity on amplitude of ELF radio
signal, Telecommun. Radio Eng., 67, No.5, 437-455, 2008.
Copy right @ESES
IJEAR
86
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Nickolaenko, A.P. and Rabinowicz, L.M., Possible global electromagnetic resonances on the planets
of the solar system, Translated from Kosmicheskie Issledovaniya, Plenum Publishing Corporation ,
20, No. 1, 82 – 88, 1982.
Nickolaenko, A.P., and Rabinowicz, L.M., Applicability of ultralow – frequency global resonances
for investigating lightning activity on Venus, Translated from Kosmicheskie Issledovaniya, Plenum
Publishing Corporation, 25, No. 2, 301 – 306, 1987.
Nickolaenko, A.P., Galuk, Yu.P., and Hayakawa, M. , Vertical profile of atmospheric conductivity
that matches Schumann resonance observations, Springer Plus, 5, 108, DOI 10.1186/s40064-0161742-3, 2016.
Nickolaenko, A.P., Hayakawa, M., and Sekiguchi, M., Variations in global thunderstorm activity
inferred from the OTD records, Geophys. Res. Lett., 33, L06823, doi:10.1029/ 2005GL024884,
2006a.
Nickolaenko, A.P., Hayakawa, M., Ogawa, T., and Komatsu, M., Q-bursts: A comparison of
experimental and computed ELF waveforms, Radio Sci., 43, RS4014, doi:10.1029/2008RS003838,
2008.
Nickolaenko, A.P., Kudintseva, I. G., Pechony, O., Hayakawa, M., Hobara, Y., and Tanaka, Y. T.,
The effect of a gamma ray flare on Schumann resonances, Ann. Geophys., 30, 1321–1329, 2012
doi:10.5194/angeo-30-1321-2012, 2012.
Nickolaenko, A.P., Rabinowicz, L.M., and Shvets, A.V., Polar non-uniformity of ionosphere related
to solar proton events, Telecommun. Radio Eng., 67(5), 413–435, 2008.
Ogawa, T. and Komatsu, M., Analysis of Q burst waveforms, Radio Sci., 42, RS2S18,
doi:10.1039/2006RS003493, 2007.
Ogawa, T., Tanaka, Y. and Yasuhara, M. , Schumann resonances and worldwide thunderstorm
activity, in Planetary Electrodynamics, Ed. by S. C. Coroniti and J. Hughes, Vol. 2, Gordon and
Breach, New York, Chap. V-7, 1969.
Ogawa, T., Tanaka, Y., Miura, T., and Yasuhara, M., Observations of natural ELF and VLF
electromagnetic noises by using ball antennas, J. Geomagn. Geoelectr., 18, 443 – 454, 1966.
Ohya, H., Shiokawa, K., and Miyoshi, Y., Development of an automatic procedure to estimate the
reflection height of tweek atmospherics, Earth, Planets and Space, 60, 837–843, 2008.
Outsu, J., Numerical study of tweeks based on wave-guide mode theory, Proc. Res. Inst. Atmos.
Nagoya Univ., 7, 58-71, 1960.
Pechony, O., Modeling and simulations of Schumann resonance parameters observed at the Mitzpe
Ramon field station (Study of the day-night asymmetry influence on Schumann resonance amplitude
records), Ph.D. Thesis, Tel-Aviv University, Israel, 92 , 2007.
Copy right @ESES
IJEAR
87
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Pechony, O. and Price, C. , Schumann resonance parameters calculated with a partially uniform knee
model on Earth, Venus, Mars, and Titan, Radio Sci., 39, RS5007, doi:10.1029/2004RS003056, 2004.
Pechony, O., and Price, C., Schumann resonances: Interpretation of local diurnal intensity
modulations, Radio Sci., 41, RS2S05, doi:10.1029/2006RS003455, 2006.
Pechony, O., Price, C. and Nickolaenko, A.P., Relative importance of the day-night asymmetry in
Schumann resonance amplitude records, Radio Sci., 42, RS2S06, doi:10.1029/2006RS003456, 2007.
Polk, C., Relation of ELF noise and Schumann resonances to thunderstorm activity, Planetary
Electrodynamics, Ed. by S. C. Coronoti and J. Hughes, Vol. 2, Gordon and Breach, New York,
Chap. V-6, 5583, 1969.
Porrat, D. and Fraser-Smith, A.C., Propagation at Extremely Low Frequencies, in Encyclopedia of
Electrical and Electronics Engineering Online, Ed. J. Webster, John Wiley and Sons, New York,
2001.
Potter, R.K., Analysis of audio-frequency atmospherics, Proc. IRE, 39(9), 1067–1069, 1951.
Rabinowicz, L.M., Global electromagnetic resonances in non-uniform
Earthionosphere cavity, Ph.D. Thesis, Kharkov (in Russian), 1988.
and
anisotropic
Rafalsky, V.A. , Resonance phenomena in cross section of the Earthionosphere duct and their
influence on radio wave excitation and propagation, Ph.D. Thesis, Kharkov (in Russian), 1991.
Rafalsky, V.A., Nickolaenko, A.P., and Shvets, A.V., Location of lightning discharges from a single
station, J. Geophys. Res., 100, No. D10, 20, 829-20, 838, 1995a.
Rafalsky, V.A., Shvets, A.V. and Hayakawa, M., One-site distance-finding technique for locating
lightning discharges, J. Atmos. Terr. Phys., 57, No. 11, 1255-1261, 1995b.
Rakov, V.A. and Uman, M.A., Lightning: Physics and Effects, Cambridge Univ. Press, Cambridge,
687, 2013.
Reeve, C.D. and Rycroft, M.J., The eclipsed lower ionosphere as investigated by natural very low
frequency radio signals, J. Atmos. Terr. Phys., 34, 667–672, 1972.
Rodger, C. J., Brundell, J. B., and Dowden, R. L., Location accuracy of VLF World Wide Lightning
Location (WWLL) network: Post-algorithm upgrade, Ann. Geophys., 23, 277-290, 2005.
Roldugin, V.C., Maltsev, Y.P., Petrova, G.A., and Vasiljev A.N., Decrease of the first Schumann
resonance frequency during solar proton events. J. Geophys. Res., 106, 18,555–18, 562, 2001.
Roldugin, V.C., Maltsev, Y.P., Vasiljev, A.N., Shvets, A.V., and Nickolaenko, A.P., Changes of
Schumann resonance parameters during the solar proton event of 14 July 2000, J. Geophys. Res.,
108(A3), 1103, doi:10.1029/2002JA009495, 2003.
Copy right @ESES
IJEAR
88
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Roldugin, V.C., Maltsev, Yu.P., Vasiljev, A.N., and Vashenyuk, E.V., Changes of the first
Schumann resonance frequency during relativistic solar proton precipitation in the 6 November 1997
event. Ann Geophysicae, 17, 1293–1297, 1999.
Ryabov, B.S., Tweek formation peculiarities. Geomagn. Aeron., English Translation. 34, no. 1, 60 –
66, 1994.
Rycroft, M.J., Harrison, R.G., Nicoll, K.A., and Mareev, E.A., An overview of Earth‟s global
electric circuit and atmospheric conductivity, in Planetary Atmospheric Electricity, Ed. by F.
Leblanc et al., Springer, New York, 83-105, doi:10.1007/978-0-387-87664-1_6.2000a, 2008.
Rycroft, M.J., Israelsson, S., and Price, C., The global atmospheric electric circuit, solar activity and
climate change, J. Atmos. Solar-Terr. Phys., 62, 1,563-1,576, 2000b.
Sátori, G., On the dynamics of the north-south seasonal migration of global lightning, Proc. 12th
ICAE, Global Lightning and Climate, Versailles, France,. 1 – 4, 2003.
Sátori, G., Mushtak., V., and Williams, E., Schumann resonance signatures of global lightning
activity. In: Betz H.D., Schumann U., Laroche P. (Eds) Lightning: Principles, Instruments and
Applications, Springer, Dordrecht, ISBN: 978-1-4020-9078-3, 347–386, 2009.
Sátori, G., Neska, M., Williams, E., and Szendröi, J., Signatures of the day-night asymmetry of the
Earth-ionosphere cavity in high time resolution Schumann resonance records, Radio Sci., 42,
RS2S10, doi:10.1029/2006RS003483, 2006.
Sátori, G., Szendröi, J., and Verö, J., Monitoring Schumann resonances  I. Methodology, J. Atmos.
Terr. Phys., 58, 14751482, 1996.
Sátori, G., Williams, E. R., and Boccippio D. J. , Global lightning and climate, AGU Fall Meeting,
San Francisco, December 8–12. 2003. P. AE32A-0166, 2003.
Sátori, G., Williams, E., Price, C., Boldi, R., Koloskov, A., Yampolski, Yu., Guha, A., and Barta, V.
, Effects of energetic solar emissions on the Earth -ionosphere cavity of Schumann resonances, Surv.
Geophys., DOI:10.1007/s10712-016-9369-z, 2016.
Sátori, G., Williams, E.R., and Mushtak, V. , Response of the Earth ionosphere cavity resonator to
the 11-year solar cycle in X-radiation, J. Atmos. Solar-Terr. Phys., 67, 553-562.
doi:10.1016/j.jastp.2004.12.006, 2005.
Schekotov, A. Yu., Nickolaenko, A.P., Hayakawa, M., Hobara, Y., Satori, G., Bor, J., and Neska, M.
, Worldwide detection of ELF transient associated with the gamma flare of December 27, 2004,
Telecommun. Radio Eng., 72 (18), 1,695-1,718, doi:10.1615/TelecomRadEng.v72.i18.60, 2013.
Schumann, W.O. , On the free oscillations of a conducting sphere which is surrounded by an air
layer and an ionosphere shell, Z. Naturforschaftung, 7a, 149-154 (in German), 1952.
Copy right @ESES
IJEAR
89
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Sekiguchi, M., Hayakawa, M., Nickolaenko, A. P., and Hobara, Y., Evidence of a link between the
intensity of Schumann resonance and global surface temperature, Ann. Geophysicae, 24, 809 – 817,
2006.
Sentman, D.D., Electrical conductivity of Jupiter shallow interior and the formation of a resonant
planetary-ionospheric cavity, ICARUS, 88, 73–86, 1990.
Sentman, D.D., Schumann Resonances, in Handbook of Atmospheric Electrodynamics, Ed. by H.
Volland, Vol.1, CRC Press, Boca Raton, FL, 267298, 1995.
Sentman, D.D. and Fraser, B.J., Simultaneous observation of Schumann resonances in California
and Australia: evidence for intensity modulation by local height of D region, J. Geophys. Res.,
96(A9), 15,973-15,984, 1991.
Sheddy, C. H., A general analytic solution for reflection from a sharply bounded anisotropic
ionosphere, Radio Sci., 3, No. 8, 792-795, 1968.
Shvets, A., and Hayakawa, M., Global lightning activity on the basis of inversions of natural ELF
electromagnetic data observed at multiple stations around the world, Surv. Geophys., 32(6), 705–
732, doi:10.1007s/10712-011-9135-1, 2011.
Shvets, A.V., Experimental study of propagation of ELFVLF radio waves and dynamics of the
global thunderstorm activity, Ph.D. Thesis, Kharkov (in Russian), 1994.
Shvets, A.V. and Hayakawa, M. , Polarization effects for tweek propagation, J. Atmos. Solar-Terr.
Phys., 60, 461–469, 1998.
Shvets, A.V., Hobara, Y., and Hayakawa, M. , Variations of the global lightning distribution
revealed from three station Schumann resonance measurements, J. Geophys. Res., 115, A12316.
doi:10.1029/2010JA015851, 2010.
Shvets, A.V., Nickolaenko, A. P., and Chebrov, V., Effect of solar flares on Schumann resonance
frequencies, report B-4, Abstracts of MMSMW-2016, Kharkov, Ukraine, 2016b.
Shvets, A.V., Nickolaenko, A. P., Chebrov, V. N., Influence of solar flares on the Schumann
resonance frequencies, Radiofizika, Izv. VUZo.(in print), 2016a.
Shvets, A.V., Serdiuk, T.M., Gorishnyaya, Y.V., Hobara, Y., and Hayakawa, M., Estimating the
lower ionosphere height and lightning location using multimode "tweek" atmospherics, J. Atmos.
Solar-Terr. Phys., 108, 1 – 9, 2014.
Singh, A. K. and Singh, R. P., Propagational features of higher harmonic tweeks at low latitudes,
Earth, Moon and Planets, 73, 277–290, 1996.
Sukhorukov, A.I., ELF-VLF atmospheric waveforms under night-time ionospheric conditions, Ann.
Geophys., 14, 33-41, 1996.
Sukhorukov, A.I., Shimakura, S. and Hayakawa, M., On the additional dispersion of a whistler in the
earth-ionosphere waveguide, Planet. Space Sci., 40(9), 1185-1191, 1992.
Copy right @ESES
IJEAR
90
International Journal of Electronics and Applied Research (IJEAR), Vol.3 Issue 2, December, 2016
Published online (http://eses.co.in/online_journal.html)
ISSN 2395 0064
Sukhorukov, A.I., Shimakura, S. and Hayakawa, M., Approximate solution for the VLF
eigenvalues near cutoff frequencies in the nocturnal inhomogeneous Earthionosphere
waveguide, Planet. Space Sci., 40, 1,3631,369, 1992.
Surkov, V. and Hayakawa, M., Ultra and Extremely Low Frequency Electromagnetic Fields,
Springer, Tokyo, 486, 2014.
Tanaka, Y. T., Hayakawa, M., Hobara, Y., Nickolaenko, A. P., Yamashita, K., Sato, M., Takahashi,
Y., Terasawa, T., and Takahashi, T., Detection of transient ELF emission caused by the extremely
intense cosmic gamma-ray flare of 27 December 2004, Geophys. Res. Lett., 38, L08805,
doi:10.1029/2011GL047008, 2011.
Toledo-Redondo, S., Salinas, A., Morente-Molinera, J.A., Mendez, A, Fornieles, J., Portí, J., and
Morente, J.A., Parallel 3D-TLM algorithm for simulation of the Earth-ionosphere cavity, J.
Computational Phys., 236, 367–379, 2013.
Velikhov, E. P., Zhamaletdinov, A. A., Sobvhakov, L. A., Veshev, A. V., Sarayev, A. K., Tokarev,
A. D., Shevtsov, A. N., Vasilyev , A. V., Sinnikov, A. G. and Yakovlev , A. V., Extra-low frequency
sounding of the Earth‟s crust with a high-power antenna, Trans. Russian Acad. Sciences (Doklady),
341, 12-16, 1996.
Volland, H., Longwave sferic propagation within the atmospheric waveguide, in Handbook of
Atmospherics, Ed. by H. Volland, Vol. 2, CRC Press, Boca Raton, FL, Chap. 3, 65–93, 1995.
Wait, J. R., Electromagnetic Waves in Stratified Media, Pergamon Press, New York, 608, 1970.
Williams, E.R., The Schumann resonance: a global tropical thermometer, Science, 256, 1,184–1,188,
1992.
Yang, H. and Pasko, V.P., Three-dimensional finite-difference time domain modeling of the Earthionosphere cavity resonances, Geophys. Res. Lett., 32, L03114. doi:10.1029/2004GL021343, 2005.
Yano, M., Ida, Y., Hayakawa, M. and Nickolaenko, A. P., Reception of ELF transmitter signals at
Moshiri, Japan and their propagation characteristics, Radio Sci., 45, RS1009, doi:10.2009RS004224,
2010.
Yano, S., Ogawa, T., and Hagino, H., Wave-form analysis of tweek atmospherics, Res. Lett. Atmos.
Electr., 9, 31–42, 1989.
Yatsevich E.I., Shvets, A. V., and Nickolaenko, A. P., Impact of ELF receiver on characteristics of
ELF transients, Radiophysics and Quantum Electronics, 57, No.3, 176 – 186, 2014.
Yedemsky, D.Ye., Ryabov, B.S., Shchokotov, A.Yu., and Yarotsky, V.S. , Experimental
investigation of the tweek field structure, Adv. Space Res., 12, 251–254, 1992.
Zhou, H.J., Hayakawa, M., Galuk, Yu.P., and Nickolaenko, A.P. , Conductivity profiles
corresponding to the knee model and relevant SR spectra, Sun and Geosphere, 11, No.1, 5–15, 2016.
Copy right @ESES
IJEAR
91